Science of the Total Environment 443 (2013) 315323

Contents lists available at SciVerse ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Phenotypic antibiotic resistance of Escherichia coli and E. coli O157 isolated from
water, sediment and biolms in an agricultural watershed in British Columbia
Rasha Maal-Bared a,, Karen H. Bartlett b, 1, William R. Bowie c, 2, Eric R. Hall d, 3
a

Resource Management and Environmental Studies, University of British Columbia, Vancouver, Canada
School of Environment and Health, 2206 East Mall, University of British Columbia, Vancouver, BC, Canada V6Z 1Z3
University of British Columbia, Division of Infectious Diseases, 452D Heather Pavilion East, Vancouver General Hospital, 2733 Heather Street, Vancouver, BC, Canada V5Z 3J5
d
Department of Civil Engineering, 6250 Applied Science Lane, The University of British Columbia, Vancouver, BC, Canada V6Z 1Z4
b
c

H I G H L I G H T S

Antibiotic resistant E. coli and E. coli O157 were prevalent in Elk Creek.
Resistance was highest to tetracycline, ampicillin and streptomycin.
Biolm and sediment isolates were more likely to be resistant than water isolates.
Some water quality variables (nutrients, temperature, DO and salinity) impact AMR.
Biolms and sediment are important reservoirs of AMR E. coli that could pose a risk to public health.

a r t i c l e

i n f o

Article history:
Received 10 August 2012
Received in revised form 25 October 2012
Accepted 25 October 2012
Available online xxxx
Keywords:
Biolms
Sediment
Minimum inhibitory concentration (MIC)
E. coli O157
Multiple antibiotic resistance (MAR)

a b s t r a c t
This study examined the distribution of antibiotic resistant Escherichia coli and E. coli O157 isolated from
water, sediment and biolms in an intensive agricultural watershed (Elk Creek, British Columbia) between
2005 and 2007. It also examined physical and chemical water parameters associated with antibiotic resistance. Broth microdilution techniques were used to determine minimum inhibitory concentrations (MIC)
for E. coli (n = 214) and E. coli O157 (n = 27) recovered isolates for ampicillin, cefotaxime, ciprooxacin,
nalidixic acid, streptomycin and tetracycline. Both E. coli and E. coli O157 isolates showed highest frequency
of resistance to tetracycline, ampicillin, streptomycin and nalidixic acid; respectively. For E. coli, the highest
frequency of resistance was observed at the most agriculturally-impacted site, while the lowest frequency
of resistance was found at the headwaters. Sediment and river rock biolms were the most likely to be associated with resistant E. coli, while water was the least likely. While seasonality (wet versus dry) had no
relationship with resistance frequency, length of biolm colonization of the substratum in the aquatic environment only affected resistance frequency to nalidixic acid and tetracycline. Multivariate logistic regressions
showed that water depth, nutrient concentrations, temperature, dissolved oxygen and salinity had statistically signicant associations with frequency of E. coli resistance to nalidixic acid, streptomycin, ampicillin
and tetracycline. The results indicate that antibiotic resistant E. coli and E. coli O157 were prevalent in an
agricultural stream. Since E. coli is adept at horizontal gene transfer and prevalent in biolms and sediment,
where ample opportunities for genetic exchange with potential environmental pathogens present themselves, resistant isolates may present a risk to ecosystem, wildlife and public health.
2012 Elsevier B.V. All rights reserved.

1. Introduction

Corresponding author at: Centre for Research on Environmental Microbiology,

University of Ottawa, 451 Smyth Road, Ottawa, Ontario, Canada K1H 8M5. Tel.: + 1
613 860 0643.
E-mail addresses: maalbared@gmail.com (R. Maal-Bared),
kbartlet@interchange.ubc.ca (K.H. Bartlett), bowie@interchange.ubc.ca (W.R. Bowie),
ehall@civil.ubc.ca (E.R. Hall).
1
Tel.: +1 604 822 6019; fax: +1 604 822 9588.
2
Tel.: +1 604 875 4147; fax: +1 604 875 4013.
3
Tel.: +1 604 822 2707; fax: +1 604 822 6901.
0048-9697/$ see front matter 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.scitotenv.2012.10.106

The increase in antimicrobial resistance (AMR) of microbial populations is an issue of global concern (Kmmerer, 2004). Antibiotics are
extensively used in agriculture to prevent and treat disease, as well as
to promote growth (Martinez, 2009; Rooklidge, 2004). Consequently,
animal waste produced in agricultural settings can contain resistant bacteria, resistance determinants present in gene-transfer units, or combinations of both (Taylor et al., 2011; Martinez, 2009; Diarra et al., 2007;
Witte, 2000). Animal waste can also contain some concentrations of antimicrobial agents that can drive selective pressures and affect microbial

316

R. Maal-Bared et al. / Science of the Total Environment 443 (2013) 315323

communities in the external environment (Martnez-Carballo et al.,

2007; Hamscher et al., 2002; Hirsch et al., 1999).
When manure produced in agriculture is applied to land, pollutants
such as antimicrobial compounds, resistant bacteria or resistance genes
concentrate and mobilize in soil and often end up in ground or surface
water through runoff (Ribeiro et al., 2012; Duriez and Topp, 2007;
Pedersen et al., 2003; Chee-Sanford et al., 2001; Hirsch et al., 1999).
Thus, it is not surprising that many studies have found antibiotic resistant bacteria in aquatic environments (Ribeiro et al., 2012; Lupo et al.,
2012; Schwartz et al., 2003; Sayah et al., 2005; Laroche et al., 2010;
Ash et al., 2002; Watkinson et al., 2007). Contamination of pristine
ecosystems with antibiotic resistant genes found in clinical settings
has been documented (Martinez, 2009; Allen et al., 2011). Aquatic
ecosystems could play an ecological and evolutionary role in driving
the persistence, emergence and spread of AMR (Taylor et al., 2011).
Understanding the distribution of antibiotic resistant bacteria in water
sources in an agricultural watershed is important since water is
the link between the four major reservoirs human, animal, soil, and
aquatic that circulate AMR (Ribeiro et al., 2012; Baquero et al.,
2008). Antibiotic-resistant genes entering into environmental ecosystems could increase risks to public health and ecologically impact population dynamics and interactions among bacterial communities
(Martinez, 2009; Rooklidge, 2004). The wide dissemination of resistance genes frequently present in human pathogens in places without
a high antibiotic load (Pallecchi et al., 2008) and the development of
resistance at very low antibiotic concentrations (Gullberg et al., 2011)
indicate that the probability of selecting for resistant bacteria in the
natural ecosystems could pose a threat to public health.
The present study was conducted to increase our knowledge of the
distribution of antibiotic resistant Escherichia coli in an agricultural
watershed. E. coli was chosen because it is a commonly used fecal
contamination indicator that is transient in aquatic environments, but
still survives well in water. It also is a gut commensal with welldocumented antibiotic resistance (Edge and Hill, 2005) of interest to a
board range of microbiologists as well, as E. coli is highly adept at horizontal gene transfer. The study was also designed to investigate the effects of a variety of physical and chemical water quality variables on
antimicrobial resistance to six antibiotics that are commonly used in
farming practices in the Elk Creek watershed (ampicillin, cefotaxime,
ciprooxacin, nalidixic acid, streptomycin and tetracycline). It is useful
to observe the effects of chemical and physical variables on bacterial
antimicrobial resistance in natural aquatic environments, as opposed
to laboratory microcosms, due to the large differences in the results of
these different types of studies (Williams et al., 1996). The working
hypothesis was that E. coli isolates retrieved from different sites and
aquatic ecosystem compartments (biolms, sediment and water) may
experience varying levels of antibiotic resistance as measured by
broth microdilution minimum inhibitory concentrations (MIC). The
MIC for a microorganism is the lowest concentration of antimicrobial
agent that will inhibit the growth of the microorganism after overnight
incubation (Andrews, 2001). The second hypothesis was that MICs
might be directly or indirectly inuenced by physical and chemical factors, such as water quality parameters and biolm colonization period.
Methods used to evaluate antibiotic resistance range from molecular
techniques to culture-based techniques. For the purposes of this
study, a culture-based technique was chosen.
2. Methods
2.1. Sampling location Elk Creek
The Elk Creek Watershed is located in the Fraser Valley to the east
of Chilliwack, British Columbia, and encompasses about 28 km 2 of
agricultural, urban, and mountainous terrain. Its headwaters are located in Elk Mountain, Mount Thurston, and the Eastern Hillsides.
Its watercourse ows north and drains into the Fraser River at Hope

Slough. The watercourse is about 12 km long (Rood and Hamilton,

1995). Peak runoff occurs in late spring during snowmelt, and
the ow becomes very low during late summer and early autumn.
The estimated mean annual ow of this watershed is 0.61 m 3.s 1
(Rood and Hamilton, 1995). The dry season typically begins in April
and ends in September, while the wet season usually begins in October
and ends in March.
The Elk Creek Watershed is mainly agricultural in nature and it hosts
a variety of greenhouses and improved pasturelands. The watershed's
lowland soils are used extensively for agriculture with dairy farming
as the main activity, and the lower portions of the watershed are dominated by chicken farms and intensive horticulture operations, most
with drainage and irrigation systems (Schendel et al., 2004). The density of livestock is estimated at 2.46 animal units per hectare. It has been
estimated that an approximate 140,000 tonnes of poultry manure are
applied to the 58% of the land in the Lower Fraser Valley annually,
while commercial fertilizers, herbicides and insecticides are applied to
58%, 29%, and 12% of the land, respectively (Vingarzan et al., 2002).
Elk Creek exhibits an increasing nutrient pollution gradient as the
stream ows towards Hope Slough (Schendel et al., 2004). Four sites
were located on the stream for the present study. This would allow
the examination of a broad range of environmental variables that may
affect E. coli AMR prevalence and distribution in the aquatic ecosystem.
The rst site (Latitude 49 08 26.5 N, Longitude 121 50 0.34 W) was
located in the headwaters. Sites 2 (Latitude 49 09 42.4 N, Longitude
121 50 52.1 W), 3 (Latitude 49 09 51.5 N, Longitude 121 51 7.8
W) and 4 (Latitude 49 10 36.3 N, Longitude 121 51 7.0 W) were
located in the agricultural area.
2.2. Biolm samplers, sample collection and preparation
Substrata used for biolm colonization were river rock, slate rock,
wood, Lexan, ne grit sandpaper, and sediment. Rock, wood and
sediment were chosen due to their natural presence in the watershed.
Sediment in the hillside around the headwaters consists of shallow
soils comprised of glacial, colluvial, and eolian sediments over bedrock. The lowland area sediment consists of mineral soils of silt
loam and silty clay loam textures, intermixed with organic soils
(Schendel et al., 2004). Hydrophobic Lexan and sandpaper were
chosen due to their prior successful use as biolm substrata by
other researchers (Hunt and Parry, 1998; Cerca et al., 2005). Biolm
samplers were built using the method described in Maal-Bared et al.
(2012). Briey, all autoclaved substrata, with the exception of sediment, were attached in duplicate to one galvanized steel bar per site
for each sampling. One bar at a time was placed at each site on the
surface of the sediment and attached to a xed object (bridge, tree,
etc.) to prevent sampler loss. The samplers were placed so that the
lengths of the galvanized steel bars were parallel to stream ow.
Samples were collected between December 2005 and April 2007
from the four sites on a rotating basis, from water, sediment and biolm
samplers. Samples collected between December 2005 and December
2006 were allowed to colonize in situ for four weeks before collection
and analysis, since previous investigations have shown that a mature
biolm community requires somewhere between 2 weeks and one
month for formation (Hunt and Parry, 1998). During this time, only
one site at a time was visited. In the wet season (October to April), 11
samplers were retrieved, while only six samplers were left in place
and retrieved in the dry season (MaySeptember). At the time of
every sampler collection, duplicate water column grab samples of
river water from the same site were collected into sterilized 500 mL
bottles, and duplicate 100 g surface sediment samples were also
taken. The surface sediment grab samples characterized the top 10
15 cm of the sediment and were collected using a standard Teon pot
that was attached to an extendable pole. Biolms collected between
January 2007 and April 2007 were allowed to colonize the substrata
for different periods of time (1, 2, 3, 4, 12 and 24 weeks). At the time

R. Maal-Bared et al. / Science of the Total Environment 443 (2013) 315323

of sampler collection, water and sediment samples were collected as

well. Throughout the study, all samples were maintained at 4 C until
the lab was reached. Since these samples were from a non-potable
source and not analyzed for compliance purposes, all samples were
processed within 24 h from collection time, as per Standard methods
for the examination of water and wastewater (APHA et al., 1998).
In the laboratory, to detach biolms from substrata, substrata
were individually placed in 500 mL of sterile PBS (PBS, pH = 7.2, 9 g
NaCl/L, 0.0067 M PO4) in sterile beakers, which were placed into a
sonication bath for 30 min. The bath was continuously cooled using
ice packs (Maal-Bared et al., 2012). The resulting PBS-biolm suspension was used for analysis. Fifty grams of sediment samples was
blended in 450 mL of sterile PBS for 3 min at low power in an
autoclaved stainless steel blender, and then transferred to a sterile
bottle. Prior to analysis, the sediment suspension was shaken for
20 s to allow for resuspension and homogenization. Water was analyzed directly.
2.2.1. E. coli and E. coli O157 identication
Identication procedures followed Standard Methods (APHA et al.,
1998). An ATCC control culture (E. coli ATCC 25922) was used as a
growth and sterility control. Identication of E. coli was performed
by incubating samples in EC broth (Oxoid Basingstoke, Hampshire,
UK) at 44.5 C and checking for gas production at 24 h and 48 h.
Positive test tubes for E. coli were conrmed by selective isolation
onto Levine's EMB plates (BBL BD Microbiological Systems, Sparks,
MD). Colonies isolated from EMB plates were then puried onto
tryptic soy agar (Difco BD Microbiology Systems, Sparks, MD) and
tested using the API 20E kit (bioMerieux , Hazelwood, MO). Conrmed E. coli colonies were also tested using the pathogenic E. coli
O157 Oxoid Dryspot test kit (Oxoid, Basingstoke, Hampshire,
UK, DR0120). After isolation and identication, samples were kept
in 5 mL glass vials in a sterile 50% Glycerol-TSA solution (Difco,
Becton, Dickinson & Co., Sparks, MD) in a 80 C freezer until antibiotic resistance testing was performed.
2.3. Broth microdilution MIC
The National Committee for Clinical Laboratory Standards (NCCLS)
standard methods were used to perform broth microdilution testing
(NCCLS, 2000). Due to the high likelihood of isolating resistant E. coli
clones from the same biolm substrata and as a result of using liquid
media in the MPN step, solutions of only one isolate per substratum
per site were prepared by culturing frozen isolates onto TSB (Difco,
Becton, Dickinson & Co., Sparks, MD). Pure cultures were suspended
in 4 mL of 0.85% sterile saline solution (nal concentration:
5 107 CFU.mL1) equivalent to a 0.5 McFarland turbidity level
(bioMrieux Vitek, Durham, NC).
Antibiotic stock solutions were prepared using the highest desired
concentration needed for testing for the following antibiotics (all from
Sigma-Aldrich Inc, St Louis, USA): ampicillin, cefotaxime, ciprooxacin,
nalidixic acid, streptomycin, and tetracycline. These antibiotics were
selected because AMR to these antibiotics has been reported in E. coli
isolated from broiler chicken in the watershed in a survey conducted
by the British Columbia Ministry of Agriculture and Lands (Pritchard
et al., 2007). Ninety-six well plates were lled with MuellerHinton
Broth (BBL, Bencton, Dickinson & Co., Sparks, MD), antibiotic solution,
and the prepared E. coli suspension. The rst and second wells in each
of the 12-well rows were always the negative control (MH broth) and
the positive control (MH broth with bacterial inoculum), respectively.
Upon inoculation, wells were incubated at 35 C overnight and read
spectrophotometrically.
MIC50% (from hereon referred to as MIC unless otherwise stated)
was determined to be the lowest concentration of the antibiotic that
reduces the growth of the organism by 50% compared to the positive
control. The streptomycin MIC was not listed in the NCCLS standard

317

method, thus it was chosen from the literature (Sunde and

Norstrom, 2005). Used MIC denitions were as follows: ampicillin
(susceptible 8 g/mL; moderately susceptible = 16 g/mL; resistant 32 g/mL); cefotaxime (susceptible 8 g/mL; moderately
susceptible=1632 g/mL; resistant64 g/mL); ciprooxacin (susceptible1 g/mL; moderately susceptible=2 g/mL; resistant4 g/mL);
nalidixic acid (susceptible16 g/mL; resistant32 g/mL); streptomycin (susceptible8 g/mL; resistant16 g/mL); and tetracycline
(susceptible 4 g/mL; moderately susceptible=8 g/mL; resistant
16 g/mL).
2.4. Water quality data
Water samples were collected to obtain physical and chemical water
quality information as described in Maal-Bared et al. (2012). Physical
water quality data included: water depth (manually measured), stream
width (manually measured), water velocity (m/s), and water ow
(calculated in m3/s and result of three variables water depth width
velocity). Other physical and chemical water quality variables measured in situ using a YSI-85 instrument (Yellow Springs, OH) were
water temperature (C), conductivity (S/cm), specic conductivity
(S/cm), and dissolved oxygen (% saturation and mg/L). Chilliwack ambient rainfall measurements (collected at the Abbotsford Airport
weather station about 45 km from the watershed) were retrieved
from the Weather Network website (http://www.theweathernetwork.
com/weather/CABC0308) and were calculated as the cumulative rainfall over 7 days before the sampling date (mm). Preserved water samples were analyzed for nitrate+ nitrite (NOX), ortho-phosphate (PO4),
ammonium (NH4+) and dissolved organic carbon (DOC) upon return
to the laboratory. Briey, the same water samples were used for
nitrate+ nitrite and ortho-phosphate analysis. The samples were dispensed in duplicate to 10 mL test tubes using a 10 mL syringe with a
0.45 m pore size lter attached. The test tubes contained a drop of
0.1% phenyl mercuric acetate to preserve the samples for analysis.
Ammonium samples were ltered in the same manner as above, but
preserved using a drop of 0.5% HCl solution. Finally, dissolved organic
carbon (DOC) samples were ltered and dispensed into 30 mL glass
vials and preserved by lowering the pH of the samples below 2 using
phosphoric acid. The nutrient analyses for nitrate + nitrite, orthophosphate and ammonium were performed using a Lachat QuikChem
8000 Flow Injection Analyzer. Ortho-phosphate concentrations were
determined using Standard Method 4500-P (Flow injection analysis
for orthophosphate) (APHA et al., 1998). Ammonium concentrations
were determined using Proposed Method 4500-NH4+ H (Flow injection analysis) (APHA et al., 1998). Nitrate+ nitrite concentrations
were found using proposed Standard Method 4500-NO3-I (Cadmium
ow injection method). DOC concentrations were obtained using the
persulfate-ultraviolet oxidation method (Standard Method 5310 C)
and using a Dohrmann Phoenix 8000 TOC analyzer (APHA et al., 1998).
2.5. Statistical analyses
All statistical analyses were completed using JMP 10.0 (SAS Institute,
NC, USA), and graphs were constructed using SigmaPlot (Version 10.0,
Systat Software, Germany). Chi-square tests were used to determine
the statistical signicance of differences in susceptibility/resistance patterns and MICs of E. coli isolated from different sites and from different
substrata. To describe relationships among MICs and water quality variables, Spearman correlations and logistic regressions were used for each
antibiotic.
Logistic regressions were used instead of linear regressions, since
the goal was to predict an ordinal, non-continuous response outcome
(antimicrobial resistance measured by MICs) through a set of variables that were continuous (water quality variables). The results
were displayed in terms of chi-square values and p-values (Quinn
and Keogh, 2002). To determine whether the effect of substratum

318

R. Maal-Bared et al. / Science of the Total Environment 443 (2013) 315323

colonization period (biolm age) on resistance was signicant,

another chi-square test was used. Alpha was set at 0.05.
Spearman correlation results provided a starting point for selecting
input parameters for ordinal regression modeling (outcomes being resistant or susceptible), using the methodology described in Black et al.
(2007). Briey, autocorrelated water quality variables were removed
from the tested models. Two variables in particular exhibited high correlations with each other (Spearman > 0.8): water temperature and
ambient temperature, and conductivity and specic conductivity.
When models were run, only the variable with the stronger correlation
was retained in the model. MICs for each antibiotic were modeled separately. Logistic regression models were rst run only using variables
that showed signicant correlations with MIC followed by lack of t
tests (called best t tests in other software).
3. Results and discussion
3.1. General phenotypic antibiotic resistance patterns in the watershed
based on broth microdilution MIC results
The current study examined the frequency of resistance of 214
E. coli isolates retrieved from four sites along an agricultural creek
and within a variety of surface biolms. Twenty-seven of these 214
isolates were E. coli O157 isolates (12.6%) and resistance among
both E. coli and E. coli O157 was prevalent. Throughout this manuscript, whenever data relating to the 214 E. coli isolates is presented,
this includes the 27 E. coli O157 isolates. Many of the E. coli isolates
tested showed some levels of antibiotic resistance, as can be seen in
Table 1. Table 1 shows the number of isolates that were susceptible,
moderately susceptible or resistant to each antibiotic. The highest
frequency of resistance was to tetracycline, followed by ampicillin,
streptomycin and nalidixic acid. While no isolates were resistant to
cefotaxime, 16 isolates were moderately susceptible to the antibiotic.
The only resistant isolate of E. coli to ciprooxacin was recovered
from sandpaper biolms at site 3.
This study did not conrm whether these 27 isolates were of the
pathogenic serotype, however E. coli O157 isolates have a high probability of being pathogenic and the results may therefore have some
implication for public and animal health protection. Also, the proportion of E. coli O157 detection is comparable to other report from raw
surface water, which range between 2%27% in Canada (Public Health
Agency of Canada, 2007) and 10%23% in other settings (Ferreira da
Silva et al., 2007). Antimicrobial resistance of E. coli isolated from
water to a variety of antibiotics has been shown in many studies
(Sayah et al., 2005; Edge and Hill, 2005; Schwartz et al., 2003;
Watkinson et al., 2007) and isolation of resistant E. coli O157 from
multiple use watersheds has been reported (Hamelin et al., 2006,
2007).
Table 1
Total number and percentage of E. coli (n = 214) and E. coli O157 (n = 27) isolates from
Elk Creek, which were susceptible, moderately susceptible (moderate) or resistant to
the antibiotics tested based on broth microdilution MICs. Note: the 27 E. coli O157 isolates are included in the data presented for the 214 E. coli isolates.
Antibiotic

Isolate

Susceptible (%)

Moderate (%)

Resistant (%)

Ampicillin

E.
E.
E.
E.
E.
E.
E.
E.
E.
E.
E.
E.

165 (77)
23 (85)
198 (93)
24 (89)
213 (99.5)
27 (100)
203 (95)
25 (93)
197 (92)
25 (93)
85 (40)
9 (33)

14 (7)
1 (4)
16 (8)
3 (11)
0
0
1
0
0
0
2 (0.9)
0

35 (16)
3 (11)
0
0
1 (0.5)
0
10 (5)
2 (7)
17 (8)
2 (7)
127 (59)
18 (67)

Cefotaxime
Ciprooxacin
Nalidixic Acid
Streptomycin
Tetracycline

coli
coli
coli
coli
coli
coli
coli
coli
coli
coli
coli
coli

O157
O157
O157
O157
O157
O157

Frequency of resistance was highest to tetracycline, followed

by ampicillin and streptomycin. High levels of E. coli antibiotic
resistance to tetracycline and ampicillin have been observed in
many other studies (Wilkerson et al., 2004; Edge and Hill, 2005;
Watkinson et al., 2007; Hlzel et al., 2010; Mudryk et al., 2010).
Edge and Hill (2005) also found low levels of resistance to ciprooxacin, which correspond with our results as well. High levels of resistance to ampicillin were expected due to fact that ampicillins are
older antibiotics that have been extensively used over the years
(Rooklidge, 2004).
While AMR in commensal microorganisms, such as E. coli, may not
seem like a risk to public health, it is important to keep several considerations in mind. E. coli is highly adept at horizontal gene transfer
and could transfer determinants of resistance to facultative or obligate pathogens that affect public health (Alexander et al., 2010).
This is particularly relevant in the case of tetracycline, irrelevant of
the source of resistance, as it remains one of the most commonly
used antibiotics in the treatment of infections in animals and humans
and as tetracycline resistance is often accompanied by resistance to
other antibiotics (Kmmerer, 2004). Studies have also shown that
the close interaction between aquatic ecosystems can result in the
migration of antibiotic resistant genes through these systems as a
result of runoff or leaching (Martinez, 2009; Laroche et al., 2010).
Thus, the presence of resistant genes in the environment could ultimately impact drinking water sources, domesticated animal and
wildlife health and the food chain. Since the AMR patterns we
observed in E. coli in Elk Creek were very similar to those observed
in broiler chicken and farmed animals in the watershed (Pritchard
et al., 2007), it is possible that exchanges between the animals population and the aquatic environment are already ongoing, particularly
in the case of E. coli O157 isolates. Finally, establishing resistance
baseline data in aquatic ecosystems can be valuable as it can be
used as an indicator of resistance evolution in different ecosystems
(Alexander et al., 2010; Mudryk et al., 2010).
3.2. Differences in phenotypic antibiotic resistance by site
The number of E. coli isolates resistant to one or more of the six
tested antibiotics varied by site. The numbers of E. coli isolates susceptible, moderately susceptible and resistant at site 1 were 46, 8 and 14
isolates, respectively (total of 68 isolates from site 1). The numbers of
E. coli isolates susceptible, moderately susceptible and resistant at site
2 were 23, 3 and 14, respectively (total of 40 isolated from site 2). The
numbers of E. coli isolates susceptible, moderately susceptible and resistant at site 3 were 38, 2 and 21, respectively (total of 61 isolates
from site 3). Finally, the numbers of E. coli isolates susceptible,
moderately susceptible and resistant at site 4 were 20, 5 and 20,
respectively (total of 45 isolates from site 4).
Highest percentage of resistant E. coli was isolated from site 4
(45%) followed by sites 2 and 3 (both 35%). Lowest percentage of
resistant isolates was retrieved from site 1 (21%). The results show
that while site 1 is less impacted by agriculture than the other three
sites located in the agricultural reach, it is still contaminated with resistant E. coli. E. coli isolate resistance to particular antibiotics differed
depending on the site. At site 1, the majority of E. coli isolates that
exhibited resistance were resistant to ampicillin (13 out of 69 isolates) and tetracycline (34 out of 69 isolates). Similar to the AMR pattern described in Alexander et al. (2010), isolates that were
tetracycline resistant were not ampicillin resistant, but isolates that
were ampicillin resistant were always tetracycline resistant. E. coli
isolates at sites 2 and 4 exhibited similar resistance patterns. Site 3
E. coli isolates were mostly resistant to tetracycline (data not shown).
Two chi-square tests were performed to evaluate whether the
differences in E. coli resistance levels among different sites were statistically signicant. The rst test comparing the general frequency of
E. coli resistance (susceptible, moderately susceptible and resistant)

R. Maal-Bared et al. / Science of the Total Environment 443 (2013) 315323

319

(Meng et al., 1998; Papandreou et al., 2000; Ash et al., 2002; Aslam
et al., 2009; Allen et al., 2011; Blanco et al., 2009; Mudryk et al., 2010).
The presence of resistant E. coli in less impacted environments,
such as site 1 in the headwaters, has been reported in other watersheds, and has been explained by wildlife contamination (Blanco
et al., 2009; Meays et al., 2006; Allen et al., 2011). The presence of
E. coli O157 is not as easily understood, as it is mostly associated
with livestock (Hancock et al., 1998; Seurinck et al., 2003), which
do not have access to site 1. Resistant E. coli O157 has been isolated
from other unpolluted aquatic environments (Hamelin et al., 2007)
and contamination of deer, possums, raccoons and birds' feces with
E. coli O157 (Renter et al., 2001, 2003; Shere et al., 1998; Hancock
et al., 1998) has been repeatedly reported in the literature. Studies
have suggested that this prevalence of resistant E. coli O157 is the result of AMR crossover between domesticated animals, livestock and
wildlife (Hagedorn et al., 1999; Sayah et al., 2005). However, since
AMR is a natural phenomenon and since E. coli is highly adept to
horizontal transfer, the observed resistance may be the product of
natural processes as well.
3.3. Differences in phenotypic antibiotic resistance by substratum
Fig. 1. Multiple antibiotic resistance (MAR) of E. coli (EC) and E. coli O157 (ECO)
isolates from Elk Creek (British Columbia) between December 2005 and April 2007
to ampicillin, cefotaxime, ciprooxacin, nalidixic acid, streptomycin and tetracycline
by site between December 2005 and April 2007.

among different sites was not statistically signicant (p-value= 0.449).

The second chi-square test evaluated whether the differences in MIC
50% of E. coli isolated from different sites were statistically signicant
for each antibiotic. The differences in ampicillin, cefotaxime, nalidixic
acid and tetracycline resistance at different sites were statistically signicant, with p-values of 0.001, 0.002, 0.002, and 0.041 respectively.
Differences in E. coli resistance to ciprooxacin and streptomycin from
different sites were not statistically signicant. It should be noted that
the statistical signicance in the cefotaxime resistance of isolates is
due to the high variability in MICs. However, the MICs did not exceed
the NCCLS denition of resistance and many of the isolates were only
moderately resistant to cefotaxime (10, 1, 0 and 5 isolates were moderately resistant to cefotaxime at sites 1, 2, 3, and 4, respectively).
Since many of the tested isolates exhibited multiple antibiotic
resistances, Fig. 1 depicts the frequency of multiple antibiotic resistances (MAR) of both E. coli (EC) and E. coli O157 (ECO) isolates. It
is important to note that all of the isolates that were resistant to
one antibiotic were resistant to tetracycline (with one exception),
while those resistant to two were resistant to tetracycline and ampicillin. The data presented in Fig. 1 treat moderately susceptible
isolates as susceptible. The graph shows that resistance to two antibiotics was common across all sites and highest in sites 1 and 3. Resistance to three antibiotics was exhibited only at sites 2 and 4. Finally,
resistance to four antibiotics only occurred at site 4.
While antimicrobial resistance (AMR) is a natural phenomenon that
bacteria use to protect themselves against competitors (Kmmerer,
2004), the use of antibiotics in medicine and agriculture has inuenced
the frequency and spread of antibiotic resistant bacteria in many aquatic
environments (Laroche et al., 2010; Ribeiro et al., 2012). In most cases,
development of AMR occurs in the gut of a host post-antibiotic treatment as a result of selective pressures. In that paradigm, water is mainly
the link between the four major ecosystems human, animal, soil, and
aquatic that circulate antibiotic resistance (Ribeiro et al., 2012;
Baquero et al., 2008) and antibiotic-resistant bacteria are transferred
between these ecosystems via fecal discharges (Pedersen et al., 2003;
Reinthaler et al., 2003; Laroche et al., 2010; Duriez and Topp, 2007;
Servais and Passerat, 2009; Oberl et al., 2012). Multiple antibiotic resistance of gram-negative bacteria in general, and E. coli isolates in particular, from animals, food and water has been reported in the literature

Fig. 2 shows the distribution of susceptible and resistant E. coli

(EC) and E. coli O157 (ECO) on different surface types pooled from
all sites in Elk Creek. Resistance frequency was highest in isolates
from sediment and river rock. Lexan and sandpaper biolms, as well
as water, were the substrata most likely to be associated with antibiotic susceptible E. coli. Fig. 2 also depicts the MAR of E. coli and E. coli
O157 isolates from biolms on different surfaces. When isolates were
resistant to one antibiotic it was tetracycline (with the exception of
the one E. coli isolate retrieved from sandpaper biolms from site
3); two were tetracycline and ampicillin; three were tetracycline, ampicillin and streptomycin; and four were tetracycline, ampicillin,
streptomycin and nalidixic acid. The co-occurrence of resistance to
those three antibiotics (ampicillin, tetracycline and streptomycin)
has been documented in farm animals (Alexander et al., 2010). The
only substratum where E. coli and E. coli O157 isolates were resistant
to four antibiotics was river rock. Slate rock biolm isolates of both
E. coli and E. coli O157 were resistant to three antibiotics, as were
E. coli isolates from river rock and sediment.

Fig. 2. Distribution of MAR exhibited by E. coli (EC) and E. coli O157 (ECO) isolates
from Elk Creek (British Columbia) to ampicillin, cefotaxime, ciprooxacin, nalidixic
acid, streptomycin and tetracycline by sampled substratum biolm or medium
(water or sediment) between December 2005 and April 2007.

320

R. Maal-Bared et al. / Science of the Total Environment 443 (2013) 315323

Two chi-square tests were performed to evaluate whether the differences in E. coli resistance levels among different substrata types
were statistically signicant. The rst tests evaluating general E. coli
resistance levels (susceptible, moderately susceptible and resistant)
on different substrata were not statistically signicant (p-value =
0.432). The second chi-square test evaluated the differences in
E. coli antibiotics' resistance levels for each antibiotic and each substratum separately using the actual MICs (MIC data not shown). The
differences in ampicillin, cefotaxime and tetracycline were statistically signicant, with p-values of b0.0001, 0.008 and 0.0427 respectively. Differences in E. coli resistance to ciprooxacin, nalidixic acid,
and streptomycin on different substrata were not statistically signicant. Due to the very small numbers of E. coli O157, the results were
not subjected to statistical treatment separately.
Similar to other studies conducted in beach, marine and aquaculture environments (Samuelson et al., 1992; Andersen and Sandaa,
1994; Mudryk et al., 2010) the results show that sediment was a
major reservoir of antibiotic resistant E. coli. While the ability of
biolms to avoid the effects of antibiotics better than their planktonic
counterparts has also been reported in the literature (Lejeune, 2003),
to our knowledge this is the rst study comparing prevalence of resistant E. coli and E. coli O157 on various surface biolms in situ. The sole
occurrence of isolates resistant to 3 or 4 antibiotics in biolms and
sediment is interesting and could be inuenced by the fact that
biolms are known to facilitate genetic exchanges among its bacterial
residents, which leads to improved horizontal resistance determinants transfer. Since rocks and sediment were shown to harbor
other pathogens, such as Campylobacter, in other studies in this watershed (Maal-Bared et al., 2012), this raises concerns related to the
potential of horizontal transfer between E. coli and pathogens that
impact animal and human health. Also, the occurrence of MAR on
substrata that are naturally present in the environment should encourage other microbiologists to investigate biolms more frequently
as they could present an important reservoir of resistance.
3.4. Relationships between antibiotic resistance and other water quality
parameters in the watershed
3.4.1. Spearman correlations
Spearman correlations were used to evaluate correlations between
MICs of different antibiotics and the other measured parameters.

The correlation coefcients and their respective levels of signicance

for correlations between MICs and the other measured water quality
parameters are listed in Table 2. Two more chi-square tests were
performed; the rst to determine whether substratum biolm colonization time had any effect on MICs, and the second to determine whether
the season (dry versus wet) had any effect on MICs.
The rst test showed that the colonization time of the biolm
substratum only had an effect on nalidixic acid (p = 0.006) and
tetracycline resistance (p = 0.010). For nalidixic acid, the highest
E. coli MIC levels were found when the substratum had been colonized
for 4 weeks. For tetracycline, the highest E. coli MIC levels were found
when the substratum had been colonized for 12 weeks, while the lowest were found when the substratum was colonized for 4 weeks. Since
horizontal gene transfer is facilitated in biolms, we expected to see a
positive association between substratum colonization period (biolms'
age) and frequency of AMR (i.e. the longer an isolate inhabits a biolms,
the more likely it is to develop resistance). This information is useful in
determining the risk of AMR development in aquatic environments in
the long run. The information could inform drinking water monitoring
and management decisions, particularly in environments where transfer potential to human and animal populations is high. However, since
the biolm population is transient and since older, thicker biolms in
aquatic environments are subject to sloughing events, it is not possible
to know whether the frequency of AMR observed after the 4-week period is due to the loss of resistant isolates to sloughing. The present
study also shows that seasonality had no statistically signicant effect
on antibiotic resistance. This result contradicts other studies that
found associations between leaching and runoff events, which are
often seasonal, and frequency of resistance in E. coli (Laroche et al.,
2010).
Microbial inhibitory concentrations for ampicillin only correlated
signicantly with PO4 levels in the water column (Spearman =
0.135, pb 0.05) and showed a strong but statistically insignicant trend
with ambient temperature levels (Spearman =0.121, p= 0.077).
Cefotamine MICs had no statistically signicant relationships with any
of the measured variables and were therefore not modeled in the next
section. Ciprooxacin MICs were only signicantly correlated with
water depth, DO% and ow (pb 0.05). Nalidixic acid MICs were closely
correlated with water depth, water temperature, NH4+ concentrations
and ambient temperature (p b 0.001). They also correlated with water
velocity, NOX concentrations and ow (p b 0.01) and to a weaker extent

Table 2
Spearman correlation coefcients and their respective p-values describing relationships between the MICs of tested antibiotics (AMP = ampicillin, CEF = cefotaxime, CIPRO = ciprooxacin,
NA = nalidixic acid, STREP = streptomycin, TET = tetracycline) and other measured parameters. Statistically signicant relationships (pb 0.05) are bolded.
AMP

CEF

CIPRO

NA

STREP

TET

Variable

Spearman

Prob > ||

Spearman

Prob > ||

Spearman

Prob > ||

Spearman

Prob > ||

Spearman

Prob > ||

Spearman

Prob >||

Velocity
Depth
Width
Water temp
DO (mg/L)
DO %
Cond
Sp. cond
Salinity
NOX
PO4
NH4+
DOC
Amb temp
Rainfall
Flow
CEF
CIPRO
NA
STREP
TET

0.025
0.055
0.075
0.114
0.052
0.043
0.056
0.096
0.054
0.075
0.135
0.034
0.001
0.121
0.033
0.103
0.001
0.014
0.018
0.088
0.082

0.716
0.420
0.272
0.095
0.454
0.532
0.417
0.162
0.436
0.278
0.044
0.623
0.992
0.077
0.633
0.134
0.985
0.839
0.796
0.198
0.235

0.071
0.069
0.132
0.095
0.015
0.047
0.054
0.017
0.044
0.060
0.083
0.009
0.104
0.064
0.035
0.032

0.299
0.313
0.054
0.164
0.829
0.493
0.435
0.805
0.524
0.381
0.229
0.893
0.130
0.349
0.606
0.646

0.105
0.185
0.056
0.011
0.058
0.138
0.019
0.075
0.024
0.091
0.096
0.126
0.108
0.072
0.107
0.161

0.127
0.007
0.436
0.879
0.395
0.043
0.781
0.273
0.725
0.184
0.160
0.065
0.114
0.297
0.119
0.019

0.179
0.421
0.160
0.324
0.170
0.124
0.060
0.171
0.068
0.217
0.029
0.263
0.063
0.300
0.118
0.198

0.009
b.0001
0.020
b.0001
0.013
0.070
0.382
0.012
0.320
0.001
0.675
b.0001
0.361
b.0001
0.086
0.004

0.168
0.119
0.203
0.145
0.200
0.093
0.266
0.223
0.319
0.084
0.077
0.073
0.231
0.004
0.281
0.131

0.014
0.083
0.003
0.035
0.003
0.175
b.0001
0.001
b.0001
0.219
0.264
0.287
0.001
0.949
b.0001
0.056

0.138
0.160
0.076
0.078
0.237
0.093
0.047
0.089
0.069
0.016
0.149
0.074
0.050
0.065
0.139
0.046

0.045
0.019
0.270
0.255
0.001
0.175
0.496
0.196
0.315
0.819
0.030
0.279
0.468
0.342
0.042
0.508

0.098
0.196
0.025
0.076

0.153
0.004
0.718
0.270

0.165
0.134
0.175

0.016
0.050
0.010

0.031
0.142

0.648
0.039

0.035

0.610

R. Maal-Bared et al. / Science of the Total Environment 443 (2013) 315323

with DO (mg/L), specic conductivity and biolm colonization period

(pb 0.05). Streptomycin MICs were closely correlated with conductivity,
salinity and rainfall (p b 0.001). Streptomycin MICs also correlated with
bank width, DO (mg/L) specic conductivity, and DOC concentrations
(pb 0.01) and to a weaker extent with water velocity and water temperature (pb 0.05). Finally, tetracycline MICs correlated with DO (mg/L)
(p= 0.001) and to a lesser extent with water velocity, depth, PO4
concentrations and rainfall (pb 0.05).
Statistically signicant relationships with both dissolved oxygen
variables were always negative (as oxygenation decreased, MIC measurements increased). While nalidixic acid and streptomycin MICs
exhibited negative relationships with bank width, water velocity,
and rainfall, tetracycline MICs exhibited positive correlations with
the same variables. Also, biolm colonization period was only positively signicantly correlated with nalidixic acid MICs.
Very few statistically signicant correlations between MICs of different antibiotics were observed. While ampicillin and streptomycin
did not correlate with any other antibiotic MICs, nalidixic acid MICs
exhibited a strong association with cefotaxime and ciprooxacin
MICs (p b 0.05). Tetracycline MICs, on the other hand, exhibited inverse statistically signicant relationships with ciprooxacin and
nalidixic acid MICs (p b 0.05).
3.4.2. Multivariate logistic regressions
Table 3 shows the water quality variables that were associated with
the presence of resistant E. coli to ampicillin, nalidixic acid, streptomycin and tetracycline. Logistic regression models generated for
cefotaxime and ciprooxacin resistance were not statistically
signicant. The strongest models were produced for nalidixic acid,
streptomycin, ampicillin and tetracycline resistance respectively
(pb 0.05). Resistance to each antibiotic had different associations with
water quality variables. Resistance to nalidixic acid was affected by
water temperature, dissolved oxygen levels (DO), and nitrate concentrations (NOX). Streptomycin and tetracycline resistances were impacted by salinity and depth, respectively. Finally, ampicillin resistance was
affected by orthophosphate concentrations (PO4) and ambient water
temperatures. Interestingly, while water temperature had a very strong
positive effect on nalidixic acid resistance, ambient temperature had a
negative impact on ampicillin resistance potentially indicating that ambient temperatures are a surrogate for another variable. A similar contradictory effect can also be observed with PO4 and NOX.
The discrepancy between relationships observed in the logistic regressions and Spearman correlations between water quality variables
and antibiotic resistance patterns is mainly due to the fact that the former used actual MICs as opposed to binary categories (resistant versus
susceptible) to evaluate associations. Depth, which is the determinant
of light penetration, had a strong positive relationship with tetracycline
resistance. Oxytetracyclines have been shown to degrade faster in
water at high temperatures, alkaline conditions and high light exposure
(Doi and Stoskopf, 2000). Photodegradation of cefotaxime, along with
18 other commonly used antibiotics, has also been shown by Alexy et
al. (2004) in the laboratory. Water temperature exhibited a strong positive association with nalidixic acid resistance. Williams et al. (1996)

321

also showed an increase in natural transformation rates in river biolms

on rocks with increased temperature.
Tetracycline resistance often exhibited inverse associations with
variables that streptomycin and ampicillin resistances had direct associations with, and vice versa; the only exception being the concentration
of orthophosphates. As orthophosphate concentrations were determined to highly impact ampicillin resistance in Table 3 (OR = 4.55,
p = 0.013), and since ampicillin resistance is often associated with
tetracycline resistance (Alexander et al., 2010), it is possible that the
Spearman correlations observed are a reection of which resistance
developed rst: tetracycline or ampicillin. However, despite the observation of a similar trend in the last ve rows of Table 2, where
ampicillin resistance was positively associated with tetracycline and
streptomycin resistances but tetracycline resistance was negatively,
these relationships were statistically insignicant. The statistically
negative correlations between tetracycline resistance and ciprooxacin
and nalidixic acid resistances, and the positive correlations between
nalidixic acid resistance and cefotaxime and ciprooxacin resistance
are more difcult to explain. The relationship between nalidixic acid
and tetracycline resistance may be driven by the water quality variables
that drive resistances as well. Water depth increased the likelihood of
tetracycline resistance but water depth also impacts water temperature
and dissolved oxygen levels negatively (see Table 3).
The effects of nutrient additions on AMR are not as clear in the
literature. Some studies suggest decreased AMR and increased susceptibility in biolm bacteria with the addition of nitrates provided
the biolm is mature (Borriello et al., 2004, 2006; Anderl et al.,
2003). Other studies suggest that the addition of nutrients enhances
horizontal transfer of genetic resistance elements and therefore increases resistance (Blanco et al., 2009). In our study, relationships
with nutrients depended on the antibiotic in question. Our results
also do not agree with Borriello et al. (2006) in that lower levels of
oxygen in the water increased the level of resistance. However,
caution must be taken when interpreting these results since the
dissolved oxygen and the nutrient concentrations gradients within
the biolms would be quite different from the concentrations in the
water column. Also, different E. coli isolates tend to have very different responses to environmental condition changes (Reisner et al.,
2006). Finally, biolm age (colonization period) had an effect on
nalidixic acid and tetracycline resistance patterns. The effects of biolm age on bactericidal effects and susceptibility have also been
shown by some studies (Monzon et al., 2001; Borriello et al., 2004,
2006).
The present study had a variety of limitations. Since we have no estimate of natural resistance in the E. coli population in Elk Creek (even at
the headwaters), we assumed that observed MIC levels are either the
result of external pressures or the result of the addition of already resistant bacteria to the system through runoff. That may not be the case, especially for tetracycline. With the exception of an on-farm survey in the
lower Fraser Valley conducted by the British Columbia Ministry of
Agriculture and Lands (Pritchard et al., 2007), which reported the prevalence of AMR zoonotic enteric pathogens isolated from broiler chicken,
dairy cows and swine, no studies have previously investigated the

Table 3
Water quality variables associated with the presence of E. coli resistant to ampicillin (AMP), nalidixic acid (NA), streptomycin (STREP) and tetracycline (TET) in Elk Creek (British
Columbia) from multivariable logistic regression models. Only statistically signicant parameters (p b 0.05) are listed along with their model estimate, standard error (SE), corresponding p-value, parameter odds ratio (OR), upper and lower 95% condence intervals (95% CI) and the summary of t (R2 and its corresponding p-value).
Antibiotic

Variable

Estimate

SE

p-Value

OR

95% CI

AMP

PO4
Ambient temp
Water temp
DO mg/L
NOX
Salinity
Depth

12.16
0.05
1.61
1.26
5.71
18.03
0.01

5.70
0.02
0.35
0.33
1.78
8.77
0.01

0.03
0.02
b0.0001
0.00
0.00
0.04
0.04

4.55
5.13
21.42
14.31
2.45
4.23
3.10

23.76
0.09
0.95
0.66
9.31
34.04
0.00

NA

STREP
TET

R2
1.02
0.01
2.36
1.97
10.24
2.40
0.02

p-Value

0.06

0.013

0.18

b0.0001

0.07
0.03

0.001
0.026

322

R. Maal-Bared et al. / Science of the Total Environment 443 (2013) 315323

concentrations of antibiotics in the watershed. Since we know that

growth promoting antibiotics are frequently used in agricultural operations in the watershed (Pritchard et al., 2007) and since large quantities of poultry manure are applied to the 58% of the land in the Lower
Fraser Valley annually (Vingarzan et al., 2002), we thought it reasonable to assume that some concentrations of antibiotics were reaching
the creek.
Another limitation is that the present study attempted to associate
resistant individuals with particular sites or substrata; however,
transport of E. coli was not taken into account. The water quality variables used in the present study may only have affected resistance indirectly and many of the variables examined are confounded in the
natural environment. Most importantly, the correlations between antibiotic resistant E. coli in biolms and water quality variables should
be interpreted with caution, since the water quality variables measured were snapshots, and do not represent the actual variation the
variable would exhibit over a prolonged period of time.
4. Conclusions
Both resistant E. coli and E. coli O157 were isolated from Elk Creek.
Frequency of resistance to ampicillin, cefotaxime, nalidixic acid and
tetracycline was signicantly different among sites sampled, being
highest at site 4 and lowest at site 1. Resistance to these same antibiotics (except nalidixic acid) also differed by surface biolm sampled
being highest in biolms associated with sediment and river rock,
and lowest in Lexan and sandpaper biolms and in water. While
seasonality (wet versus dry) had no relationship with resistance frequency, colonization period of the substratum only affected resistance frequency to nalidixic acid and tetracycline. Also, a variety of
water quality variables showed strong relationships with antimicrobial resistance that may be worth examining further. These variables
include water depth, nutrient concentrations, temperature, dissolved
oxygen and salinity. The results from this study could be of interest to
a wide audience, as E. coli is a microorganism that is procient at horizontal transfer and that could transmit its resistance genes to other
facultative or obligate pathogenic bacteria. To our knowledge, this is
the only study that investigates the occurrence of AMR and MAR in
biolms in a natural aquatic environment. The results are also novel
in that they improve our understanding of what water quality variables impact AMR development in aquatic environments that have
been shown to interact, thereby aiding management decisions in a
variety of environments where possibility of resistant gene transfer
is increased.
Acknowledgements
This research has been funded by the CIHR Strategic Training Program in Public Health and the Agricultural Rural Ecosystem (PHARE)
and Partner Institutes including the Institute of Cancer Research,
Institute of Circulatory and Respiratory Health, Institute of Infection
and Immunology, and the Institute of Population and Public Health.
We would also like to acknowledge the School of Environmental
Health where all the microbiological lab analyses were completed.
Finally, we would like to thank the University of British Columbia
Environmental Engineering Laboratory, particularly Paula Parkinson,
for conducting the water nutrient analyses.
References
Alexander TW, Inglis GD, Yanke LJ, Topp E, Read RR, Reuter T, et al. Farm-to fork characterization of Escherichia coli associated with feedlot cattle with a known history
of antimicrobial use. Int J Food Microbiol 2010;137:408.
Alexy R, Kumpel T, Kmmerer K. Assessment of degradation of 18 antibiotics in the
closed bottle test. Chemosphere 2004;57(6):50512.
Allen SE, Boerlin P, Janecko N, Lumsden JS, Jardine C. Antimicrobial resistance in generic
Escherichia coli isolates from wild small mammals living in swine farm, residential,

landll, and natural environments in southern Ontario, Canada. Appl Environ

Microbiol 2011;77:8828.
Anderl JN, Zahller J, Roe F, Stewart PS. Role of nutrient limitation and stationary-phase
existence in Klebsiella pneumoniae biolm resistance to ampicillin and ciprooxacin.
Antimicrob Agents Chemother 2003;47:12516.
Andersen SR, Sandaa RA. Distribution of tetracycline resistance determinants among
Gram-negative bacteria isolated from polluted and unpolluted marine sediments.
Appl Environ Microbiol 1994;60:90812.
Andrews J. Determination of minimum inhibitory concentrations. J Antimicrob
Chemother 2001;48:5-16.
APHA, AWWA, WEF. Standard methods for the examination of water and wastewater.
20th ed. Washington, DC: American Public Health Association (APHA), American
Waterworks Association (AWWA), and Water Environment Association (WEF);
1998.
Ash R, Mauck B, Morgan M. Antibiotic resistance of gram-negative bacteria in rivers,
United States. Emerg Infect Dis 2002;8(7):7136.
Aslam M, Diarra MS, Service C, Rempel H. Antimicrobial resistance genes in Escherichia
coli isolates recovered from a commercial beef processing plant. J Food Prot
2009;72(5):108993.
Baquero F, Martinez JL, Canton R. Antibiotics and antibiotic resistance in water environments. Curr Opin Biotechnol 2008;19:2605.
Black LE, Brion GM, Freitas SJ. Multivariate logistic regression for predicting total
culturable virus presence at the intake of a potable-water treatment plant: novel
application of the atypical coliform/total coliform ratio. Appl Environ Microbiol
2007;73:396574.
Blanco G, Lemus JA, Grande J. Microbial pollution in wildlife: linking agricultural
manuring and bacterial antibiotic resistance in red-billed choughs. Environ Res
2009;109(4):40512.
Borriello G, Werner E, Roe F, Kim AM, Ehrlich GD, Stewart PS. Oxygen limitation contributes to antibiotic tolerance of Pseudomonas aeruginosa in biolms. Antimicrob
Agents Chemother 2004;48:265964.
Borriello G, Richards L, Ehrlich GD, Stewart PS. Arginine or nitrate enhances antibiotic
susceptibility of Pseudomonas aeruginosa in biolms. Antimicrob Agents Chemother
2006;50:3824.
Cerca N, Pier GB, Vilanova M, Oliveira R, Azeredo J. Quantitative analysis of adhesion
and biolm formation on hydrophilic and hydrophobic surfaces of clinical isolates
of Staphylococcus epidermis. Res Microbiol 2005;15:50614.
Chee-Sanford JC, Aminov RI, Krapac IJ, Garrigues-Jeanjean N, Mackie RI. Occurrence and
diversity of tetracycline resistance genes in lagoons and groundwater underlying
two swine production facilities. Appl Environ Microbiol 2001;67:1494502.
Diarra MS, Silversides FG, Diarrassouba F, Pritchard J, Masson L, Brousseau R, et al.
Impact of feed supplementation with antimicrobial agents on growth performance
of broiler chickens, Clostridium perfringens and Enterococcus number, antibiotic
resistant phenotype, and distribution of antimicrobial resistance determinants in
Escherichia coli. Appl Environ Microbiol 2007;73(20):656676.
Doi AM, Stoskopf MK. The kinetics of oxytetracycline degradation in deionized water
under varying temperature, pH, light, substrate, and organic matter. J Aquat
Anim Health 2000;12:24653.
Duriez P, Topp E. Temporal dynamics and impact of manure storage on antibiotic resistance patterns and population structure of Escherichia coli isolates from a commercial swine farm. Appl Environ Microbiol 2007;73:548693.
Edge TA, Hill S. Occurrence of antibiotic resistance in Escherichia coli from surface
waters and faecal pollution sources near Hamilton Ontario. Can J Microbiol
2005;51:5015.
Ferreira da Silva M, Vaz-Moreira I, Gonzalez-Pajuelo M, Nunes OC, Manaia CM. Antimicrobial resistance patterns in Enterobacteriaceae isolated from an urban wastewater treatment plant. FEMS Microbiol Ecol 2007;60:16676.
Gullberg E, Cao S, Berg OG, Ilbck C, Sandegren L, Hughes D, et al. Selection of resistant
bacteria at very low antibiotic concentrations. PLoS Pathog 2011;7(7):e1002158.
http://dx.doi.org/10.1371/journal.ppat.1002158.
Hagedorn C, Robinson SL, Filtz JR, Grubbs SM, Angier TA, Reneau Jr RB. Using antibiotic
resistance patterns in the fecal streptococci to determine sources of fecal pollution
in a rural Virginia watershed. Appl Environ Microbiol 1999;65:552231.
Hamelin K, Bruant G, El-Shaarawi A, Hill S, Edge TA, Bekal S, et al. A virulence and antimicrobial resistance DNA microarray detects a high frequency of virulence genes
in Escherichia coli isolates from Great Lakes recreational waters. Appl Environ
Microbiol 2006;72:42006.
Hamelin K, Bruant G, El-Shaarawi A, Hill S, Edge TA, Fairbrother J, 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. Appl Environ Microbiol 2007;73:47784.
Hamscher G, Sczesny S, Hper H, Nau H. Determination of persistent tetracycline residues in soil fertilized with liquid manure by high-performance liquid chromatography with electrospray ionization tandem mass spectrometry. Anal Chem
2002;74:150918.
Hancock DD, Besser TE, Rice DH, Ebel ED, Herriott DE, Carpenter LV. Multiple sources of
Escherichia coli O157 in feedlots in the Northwestern USA. Prev Vet Med 1998;35:119.
Hirsch R, Ternes T, Haberer K, Kratz KL. Occurrence of antibiotics in the aquatic environment. Sci Total Environ 1999;225:10918.
Hlzel CS, Schwaiger K, Harms K, Kuchenhoff H, Kunz A, et al. Sewage sludge and liquid pig
manure as possible sources of antibiotic resistant bacteria. Environ Res 2010;110(4):
31826.
Hunt AP, Parry JD. The effect of substratum roughness and river ow rate on the development of a freshwater biolm community. Biofouling 1998;12(4):287303.
Kmmerer K. Resistance in the environment. J Antimicrob Chemother 2004;54:
31120.

R. Maal-Bared et al. / Science of the Total Environment 443 (2013) 315323

Laroche E, Petit F, Fournier M, Pawlak B. Transport of antibiotic-resistant Escherichia
coli in a public rural karst water supply. J Hydrol 2010;392:1221.
Lejeune P. Contamination of abiotic surfaces: what a colonizing bacterium sees and
how to blur it. Trends Microbiol 2003;11:17984.
Lupo A, Coyne S, Berendonk TU. Origin and evolution of antibiotic resistance: the common mechanisms of emergence and spread in water bodies. Front Microbiol
2012;3:18.
Maal-Bared R, Bartlett KH, Bowie WR, Hall ER. Campylobacter spp. distribution in
biolms on different surfaces in an agricultural watershed (Elk Creek, British
Columbia): using biolms to monitor for Campylobacter. Int J Hyg Environ Health
2012;215(3):2708.
Martinez JL. Environmental pollution by antibiotics and by antibiotic resistance determinants. Environ Pollut 2009;157:2893902.
Martnez-Carballo E, Gonzlez-Barreiro C, Scharf S, Gans O. Environmental monitoring
study of selected veterinary antibiotics in animal manure and soils in Austria.
Environ Pollut 2007;148(2):5709.
Meays CL, Broersma K, Nordin R, Mazumder A, Samadpour M. Spatial and annual
variability in concentrations and sources of Escherichia coli in multiple watersheds.
Environ Sci Technol 2006;40:528996.
Meng J, Zhao S, Doyle MP, Joseph SW. Antibiotic resistance of Escherichia coli O157:H7
and O157:NM isolated from animals, food, and humans. J Food Prot 1998;61(11):
15114.
Monzon M, Oteiza C, Leiva J, Amorena B. Synergy of different antibiotic combinations in
biolms of Staphylococcus epidermidis. J Antimicrob Chemother 2001;48:793801.
Mudryk Z, Perlinski P, Skorczewski P. Detection of antibiotic resistant bacteria inhabiting
the sand of non-recreational marine beach. Mar Pollut Bull 2010;60(2):20714.
National Committee for Clinical Laboratory Standards. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobicallysixth edition: approved
standard M7-A5. Wayne, PA, USA: NCCLS; 2000.
Oberl K, Capdeville MJ, Berthe T, Budzinski H, Petit F. Evidence for a complex relationship between antibiotics and antibiotic-resistant Escherichia coli: from medical
center patients to a receiving environment. Environ Sci Technol 2012;46:185968.
Pallecchi L, Bartoloni A, Paradisi F, Rossolini GM. Antibiotic resistance in the absence of
antimicrobial use: mechanisms and implications. Expert Rev Anti Infect Ther
2008;6(5):72532.
Papandreou S, Pagonopoulou O, Vantarakis A, Papapetropoulou M. Multiantibiotic resistance of gram-negative bacteria isolated from drinking water samples in southwest Greece. J Chemother 2000;12(4):26773.
Pedersen J, Yeager M, Suffet I. Xenobiotic organic compounds in runoff from elds irrigated with treated wastewater. J Agric Food Chem 2003;51:136072.
Pritchard J, Wetzstein M, DeWith N, Byrne S, Zabek E. On-farm baseline surveillance for
zoonotic enteric pathogens within the lower Fraser River valley. BC Ministry of Agriculture and Lands. Presented at Animal Determinants of Emerging Disease
(ADED) Rounds at Centre for Coastal Health on March 20; 2007. (Retrieved June 29,
2012. Available at: http://www.google.ca/url?sa=t&rct=j&q=&esrc=s&source=
web&cd=1&ved=0CEwQFjAA&url=http%3A%2F%2Fwww.viu.ca%2Fcch%2Faded%
2Fdocuments%2FPritchardMarch2007ADED.ppt&ei=MrDsT8yxKuT50gGyg_3BQ&usg=AFQjCNGzO1HSfwKGCpntoVDwb2Wm6wVVzg&sig2=7vbV1t0InySHpCqLqI
QxAg).
Public Health Agency of Canada (PHAC). C-Enternet 2007 annual report: National Integrated Enteric Pathogen Surveillance Program. Ottawa: Public Health Agency of
Canada; 2007. (Retrieved on August 9, 2012. Available at: http://www.phac-aspc.
gc.ca/publicat/2010/c-enternet07/index-eng.php#toc).
Quinn GP, Keogh MJ. Experimental design and data analysis for biologists. Cambridge,
England: Cambridge University Press; 2002.

323

Reinthaler FF, Posch J, Feierl G, Wuest G, Haas D, Ruckenbauer G, et al. Antibiotic resistance of E. coli in sewage and sludge. Water Res 2003;37:168590.
Reisner A, Krogfelt KA, Klein BM, Zechner EL, Molin S. In vitro biolm formation of
commensal and pathogenic Escherichia coli strains: impact of environmental and
genetic factors. J Bacteriol 2006;188:357281.
Renter DG, Sargeant JM, Hyngstorm SE, Hoffman JD, Gillespie JR. Escherichia coli in
free-ranging deer in Nebraska. J Wildl Dis 2001;37:75560.
Renter DG, Sargeant JM, Oberst RD, Samadpour M. Diversity, frequency, and persistence of Escherichia coli O157 strains from range cattle environments. Appl Environ
Microbiol 2003;69:5427.
Ribeiro AF, Laroche E, Hanin G, Fournier M, Quillet L, Dupont J-P, et al. Antibiotic-resistant
Escherichia coli in karstic systems: a biological indicator of the origin of fecal contamination? FEMS Microbiol Ecol 2012;81:26780.
Rood KM, Hamilton RE. Hydrology and water use for salmon streams in the Chilliwack/
Lower Fraser habitat management area, BC. Prepared for the Fraser River Action Plan
by Northwest Hydraulic Consultants Ltd., North Vancouver, BC; 1995.
Rooklidge SJ. Environmental antimicrobial contamination from terraccumulation and
diffuse pollution pathways. Sci Total Environ 2004;325:1-13.
Samuelson OB, Torsvik V, Ervik A. Long-range changes in oxytetracycline concentration
and bacterial resistance towards oxytetracycline in sh farm sediment after
medication. Sci Total Environ 1992;114:2536.
Sayah RS, Kaneene JB, Johnson Y, Miller RA. Patterns of antimicrobial resistance observed in Escherichia coli isolates obtained from domestic and wild-animal fecal
samples, human septage, and surface water. Appl Environ Microbiol 2005;71:
1394404.
Schendel EK, Schreier H, Lavkulich LM. Linkages between phosphorus index estimates
and environmental quality variables. J Soil Water Conserv 2004;59:24351.
Schwartz T, Kohnen T, Jansen B, Obst U. Detection of antibiotic-resistant bacteria and
their resistance genes in wastewater, surface water, and drinking water biolms.
FEMS Microbiol Ecol 2003;43:32535.
Servais P, Passerat J. Antimicrobial resistance of fecal bacteria in waters of the Seine
river watershed (France). Sci Total Environ 2009;408:36572.
Seurinck S, Verstraete W, Siciliano SD. Use of 16S23S rRNA intergenic spacer region
PCR and repetitive extragenic palindromic PCR analyses of Escherichia coli isolates
to identify nonpoint fecal sources. Appl Environ Microbiol 2003;69:494250.
Shere JA, Bartlett KJ, Kaspar CW. Longitudinal study of Escherichia coli O157:H7 dissemination on four dairy farms in Wisconsin. Appl Environ Microbiol 1998;64:13909.
Sunde M, Norstrom M. The genetic background for streptomycin resistance in Escherichia
coli inuences the distribution of MICs. J Antimicrob Chemother 2005;56:8790.
Taylor NGH, Verner-Jeffreys DW, Baker-Austin C. Aquatic systems: maintaining, mixing
and mobilizing antimicrobial resistance? Trends Ecol Evol 2011;26:27884.
Vingarzan R, Belzer W, Thompson B. Nutrient levels in the atmosphere of the Elk Creek
Watershed, Chilliwack, BC (1999-2000). Environment Canada technical report.
Vancouver, Canada: Aquatic and Atmospheric Sciences Division; 2002.
Watkinson AJ, Micalizzi GB, Graham GM, Bates JB, Costanzo SD. Antibiotic-resistant
Escherichia coli in wastewaters, surface waters, and oysters from an urban riverine
system. Appl Environ Microbiol 2007;73(17):566770.
Wilkerson C, Samadpour M, van Kirk N, Roberts MC. Antibiotic resistance and distribution of tetracycline resistance genes in Escherichia coli O157:H7 isolates from
humans and bovines. Antimicrob Agents Chemother 2004;48:106677.
Williams HG, Day MJ, Fry JC, Stewart GJ. Natural transformation in river epilithon. Appl
Environ Microbiol 1996;62(8):29948.
Witte W. Ecological impact of antibiotic use in animals on different complex microora:
environment. Int J Antimicrob Agents 2000;14:3215.