Heliyon 5 (2019) e02996

Contents lists available at ScienceDirect

Heliyon
journal homepage: www.cell.com/heliyon

Research article

Prevalence and antibiotic susceptibility pattern of Escherichia coli O157:H7

isolated from harvested fish at Lake Hayq and Tekeze dam,
Northern Ethiopia
Ayalew Assefa a, *, Fikru Regassa b, Dinka Ayana c, Kebede Amenu d, Fufa Abunna e
a
Sekota Dryland Agricultural Research Center P.O. Box 62, Sekota, Ethiopia
b
Addis Ababa University, College of Veterinary Medicine and Agriculture, Department of Biomedical Sciences, P.O. Box 34, Bishoftu, Oromia, Ethiopia
c
Addis Ababa University, College of Veterinary Medicine and Agriculture, Department of Pathology and Parasitology, P.O. Box 34, Bishoftu, Oromia, Ethiopia
d
Addis Ababa University, College of Veterinary Medicine and Agriculture, Department of Microbiology, Immunology and Public Health, P.O. Box 34, Bishoftu, Oromia,
Ethiopia
e
Addis Ababa University, College of Veterinary Medicine and Agriculture, Department of Clinical Studies, P.O. Box 34, Bishoftu, Oromia, Ethiopia

A R T I C L E I N F O A B S T R A C T

Keywords: Fisheries play a significant role in food security, livelihood, and source of income in developing countries.
Food science Although fish are a healthy source of protein, they can also spread diseases caused by pathogenic micro-organisms
Microbiology they may harbor. Epidemiology of foodborne pathogens is not well studied in Ethiopia. To address this issue to
Environmental science
some extent, a cross-sectional study with a simple random sampling approach was conducted from October 2017
Veterinary science
to May 2018 with the objectives of to isolate and estimate the prevalence of Escherichia coli O157: H7 in fish, and
Fish
Antibiotic susceptibility test to evaluate the antimicrobial susceptibility pattern of the isolates in selected Lakes of Northern Ethiopia. All the
Hayq microbial identification and isolation procedures were conducted based on ISO 6887-3:2017 recommendations.
Prevalence Antimicrobial susceptibility test was also performed following the standard procedure of Kirby-Bauer disk
Tekeze diffusion protocol. From the total 410 fish samples examined, six (1.46%) of them were found contaminated with
Escherichia coli O157 Shiga toxin-producing Escherichia coli O157: H7 strain. The organism was isolated from landing sites (5/293) and
local retail markets (1/75). Besides, Escherichia coli O157: H7 was isolated from filleted fish (5/214) and whole
fish (1/125); however, it was not isolated from samples of ready to eat fish and working environments of res-
taurants. The antibiotic susceptibility test revealed that the isolates were resistant to Ampicillin and Streptomycin
disks. However, Ciprofloxacin, Gentamicin and Nalidixic acid were found effective in inhibiting the growth of all
of the isolates. Since pathogenic Escherichia coli strain was detected from fish, raw and undercooked fish con-
sumption in Ethiopia may result in contracting infections. The occurrence of such pathogenic organisms in fish
indicates the need for intervention by stakeholders. Supports like freezers, generators, the establishment of fish
processing plants and on job training about proper fish handling practices may play a tremendous role in
decreasing the level of contamination of fish in Ethiopia.

1. Introduction capita fish supply has outpaced population growth in the past five de-
cades and reached a new record high of 20 kg in 2014 due to its vigorous
Fisheries play a significant role in food security, livelihood, a source growth. In the last two decades, dramatic growth in aquaculture pro-
of income and social development in developing countries (Hossain et al., duction has boosted the average consumption of fish and fishery products
2015). New technological advances and increased demands for fish as a at the global level (FAO, 2016).
source of animal protein are the main reasons for the industry's growth. Despite being a good source of high-quality animal protein and in-
Fisheries and aquaculture remain important sources of food, nutrition, come for the poor, fish can be a source of foodborne infections and in-
income and livelihoods for hundreds of millions of people around the toxications if it is not harvested and processed under hygienic conditions.
world (Garcia-Rodríguez and De La Cruz-Aguero, 2011). World per Fish can be a vehicle for several bacterial disease transmissions (Elsaidy

* Corresponding author.
E-mail address: hayall2020@gmail.com (A. Assefa).

https://doi.org/10.1016/j.heliyon.2019.e02996
Received 23 September 2018; Received in revised form 22 October 2019; Accepted 4 December 2019
2405-8440/© 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
A. Assefa et al. Heliyon 5 (2019) e02996

et al., 2015; Zhao et al., 2018). Fish can be contaminated with pathogens 2. Material and methods
through direct contact with water as well as during processing after
harvest in the environment. Enterobacteriaceae are widely distributed in 2.1. Study areas
water bodies and can be detected in fish skin, gills and intestines (Ribeiro
et al., 2015). The unhygienic conditions of the landing sites, storage, and The study was conducted in the Amhara region in Northern Ethiopia
domestic retail markets are causing a tremendous public health risk. The within the surrounding of Tekeze human-made reservoir and Lake Hayq.
fecal contamination of natural water bodies has emerged as a challenge The reservoir is built on the Tekeze river which is a tributary of the Nile.
in developing countries. The fish harvested from such areas often contain The Dam has an average area and depth of 160.4km2 and 58m, respec-
human pathogenic microorganisms. Escherichia coli has been tradition- tively (Teame et al., 2016). It has a tremendous resource regarding fish
ally recognized as an indicator organism of fecal contamination of water production that reaches an average of 5,065 tons of fish annually which
and fish (Albuquerque, 2013). created invaluable job opportunities for the surrounding community.
Foodborne pathogens are the leading causes of foodborne human Lake Hayq is a freshwater lake located in South Wollo zone of Amhara
illness and death in the world (Agüeria et al., 2018). The reason for the region, Northern Ethiopia. It is one of the Highland Lakes of Ethiopia,
increased risk can be attributed to many reasons; changes in eating located at an altitude of 2,030 m above sea level. It receives water from
habits, mass catering, complex and lengthy food supply procedures with many small seasonal streams and one perennial river named Ankerka.
increased international movement and poor hygiene practices are few of The catchment area of the lake is 65 km2. The lakes had a maximum
them (Bevilacqua et al., 2017; Lopez-Campos et al., 2012). Escherichia coli depth of 88.2m and 81.44m recorded in 1941 and 2013, respectively
O157 is one of the virulent strains under the species Escherichia coli. It is (Seid, 2016). This lake is well known for its fishery resource for the
the leading cause of hemorrhagic colitis, hemolytic-uremic syndrome community and nearby towns. It has the potential of an average of 500
(HUS) and thrombotic thrombocytopenic purpura (TTP) in man. These tons of fish production per year (Tesfaye and Wolff, 2014). The two study
illnesses may lead to death due to improper absorption of nutrients and Lakes are illustrated in Figure 1.
the destruction of specific tissues in the target organs (Cobbaut et al.,
2009; Earley and Leonard, 2006).
2.2. Study design, sampling method, and sample size determination
Epidemiology of foodborne pathogens, especially that of pathogenic
Escherichia coli, is not well studied in Ethiopia. However, recently there is
A cross-sectional study design with a simple random sampling
an increasing trend of reporting the occurrence level of the organism in
approach was implemented from October 2017 to May 2018. Fish har-
beef and dairy products (Abebe et al., 2014; Desta et al., 2016; Hailese-
vested from the two Lakes were sampled in different locations (landing
lassie et al., 2013; Hiko et al., 2008; Tassew et al., 2010; Taye, 2013).
site, local retail market, and restaurants). In the study areas, fish is sold in
This may be because beef and dairy products are considered the risk for
a jerry can (casa) and sac containers. Before reaching restaurants, the fish
E. coli due to close contact with cattle manure. However, studies showed
stay for more than 6h in landing sites. In this location, containers (casa
that the organism had been isolated from mutton, chicken, fish, and
and sacs) were randomly selected in a simple random sampling method.
vegetables. Escherichia coli O157: H7 have been reported from fish in
On average, 20–30 casa and sacs of fish were sampled in each visit
India and USA. A study conducted in the USA showed that E. coli O157:
randomly from the total of 75–100 casa and sacs available in average.
H7 genes had been detected in a gastrointestinal sample from fish with a
From a selected jerry can or sac container, few swabs from different fish
prevalence of 6.95% (Ribeiro et al., 2015). In India, O157 genes were
were taken and pooled together into a single transport media as a single
identified from fish samples taken from the street market (Sanath Kumar
sample, which represents one sample of the total sample size. Besides, for
et al., 2001).
fish tissue samples, a container was randomly selected as well. Fish were
Fishes and fish products can be contaminated by Enterobacterial
randomly selected and approximately 20g of the muscle was cut and
species that can cause a severe foodborne outbreak (Ribeiro et al., 2015).
added to stomacher bags. In local markets, samples were collected with a
The risk can be worse in developing countries like Ethiopia due to many
simple random sampling approach, as well. All restaurants were included
reasons which include but not limited to; lack of awareness of fishermen
in the study sites of Hayq city because of the absence of enough restau-
about food-borne infections or the consequences of non-hygienic
rants that serve fish. From each restaurant, a single working environment
handling of fish, extensive cattle production around water bodies and
sample. However, from these restaurants, fish samples and working
landing sites, suitable weather condition for, extensive cattle production
environment samples were also taken in a simple random sampling
around water bodies and landing sites, suitable weather condition for the
approach. Workers and processed fish were randomly selected, whereas,
flourishing of microorganism in feces, widespread open defecation by
from working knives and containers, swabs were pooled together as a
rural population, high flooding that drain contaminants to Lakes and
single sample.
landing sites, and unavailability of fish processing plants can be few of
The prevalence of E. coli O157: H7 in fish, the sample size was
them. All those listed reasons indicate that consumption of raw and
determined by (Thrusfield, 2007) as follows;
undercooked fish and fish products in Ethiopia should not be practiced.
As it is well known contaminated water (contaminated with ruminant 1:962*pexp ð1  pexpÞ
feces) is one of the most important source of E. coli O157:H7, therefor any n¼
d2
study dealing with the detection of E. coli O157:H7 from lake water
samples in the country primarily in study areas (Tekeze & Lake Hayq). where n, is the required total sample size, pexp is the expected prevalence
Even though fish pose a significant risk of foodborne pathogens, to the of the organism in fish and d, stands for the desired absolute precision of
best of our knowledge, there is no single study conducted in Ethiopia 5%. Because there was no prior information regarding the prevalence
reporting Escherichia coli O157:H7 associated with fish. In addition to being estimated, a more conservative expectation of 50% prevalence of
problems related to poor handling practices of fish and fish products, Escherichia coli O157: H7 was considered. Hence, the sample size was
post-harvest loss and causes are not yet studied in the study areas. If estimated to be a minimum of 384 fish as well as other expected sources
fishermen and policymakers are being able to make informed decisions in of contamination samples from the landing sites and restaurants, and we
an attempt to reduce contamination and losses, the significance of these sampled 410 fish and working environment samples from the two Lakes.
problems need to be scientifically studied, and recommendations should 250 samples were taken from lake Hayq, while the 160 samples were
be generated. Bearing this idea in mind, the study was designed with the from Tekeze dam. This sample size imbalance was due to the absence of
objectives of assessing the level of contamination of fish with Shiga toxin- local retail markets and restaurants around Tekeze dam. Samples from
producing Escherichia coli O157: H7 and determining in vitro suscepti- local retail markets and restaurants were taken only from Lake Hayq and
bility pattern of the isolates to commercially available antibiotics. nearby Hayq city.

2
A. Assefa et al. Heliyon 5 (2019) e02996

Figure 1. A map that shows study sites.

2.3. Source of fish and sample collection procedure The final serological confirmation of positive samples was not carried
out in the above laboratories. For that matter, isolates were prepared for
The source of the fish used in the study were collected from fishers transportation to Addis Ababa University-college of veterinary medicine
operating in the two study lakes. and agriculture (AAU-CVMA) laboratory as follows; isolates from the
Filleted fish swab, filleted fish muscle (tissue), whole fish (skin) swab, nutrient media were sub-cultured onto Trypton soya broth (Oxoid, Ltd.,
knife and cutting board swab, ready to eat fish from restaurants, workers Basingstoke, UK) for 24 h at 37  C. One milliliter of the Trypton soya
hand swab, and container swab were collected by following recom- broth (TSB) bacterial suspension was mixed with an equal volume of
mended procedures (ISO6887-3:2017, 2017). Approximately 20g of fish sterilized 50% glycerol in sterilized Cryovial tubes. These bacterial-
was cut and inserted into a stomacher bag that contained 200ml of glycerol stocks were stored at -80  C deep freezer and transported with
transport media (buffered peptone water). Besides, swabs from the whole icebox packed with ice to AAU-CVMA lab. Samples were stored in a deep
fish by swabbing the almost all the surface of the fish, a hand of the freezer and when Latex kit is secured, bacterial stocks were revived by
worker, container and, a knife was collected in a universal bottle that culturing onto TSB and further subculturing onto nutrient agar. From
contains 10ml buffered peptone water as a transport media. Samples nutrient, agar colonies were further subjected to biochemical tests.
collected from the respective sites were transported to the nearest labo- Indole test and dextrose/lactose fermentation test by culturing onto TSB
ratory in the shortest possible time and processed upon arrival. and streaking onto Klinger iron agar (KIA) media respectively, were
carried out to screen positive samples. Based on these tests, red ring
2.4. Sample preparation and identification procedure of Escherichia coli formation by indole test, and yellow slant and butt formation for the
O157:H7 sugar fermentation test was presumed to be positive for Escherichia coli
O157: H7. Positive cultures were further subjected to a serological test
All the isolation and identification procedures of the organism were using a latex kit for the confirmation of E. coli O157: H7 strain. The
performed based on ISO 6887-3:2017 recommendation for microbio- bacterial colony was picked and subjected to a slide agglutination test
logical analysis of fish samples (ISO6887-3:2017, 2017). Samples using an E. coli O157 latex kit (Oxoid Ltd., Hampshire, UK). A drop of test
collected from study sites were transferred appropriately to the nearby latex and sterile saline water was dispensed into the reaction card
animal health laboratories (Sekota Dryland Agricultural Research Center separately. Up to five presumptive E. coli O157: H7 colonies were picked
for samples taken from Tekeze dam, and Kombolcha Regional Animal by lightly touching the center of the colony with a sterile inoculating
Health Laboratory for samples taken from Lake Hayq) with a transport needle. The picked colonies were thoroughly emulsified with the saline
media. After reaching the laboratory, samples were incubated at 41.5  C on latex card and then finally with the test latex reagent. The formation of
for 6h to increase the recovery rate of stressed cells. Muscle samples of agglutination within 1 min was regarded as positive.
fish placed in a plastic bag were homogenized using a homogenizer. From
each incubated suspension (swabs in 10ml of BPW, and homogenized 2.5. Antibiotic susceptibility test
muscle samples) approximately, 200μl of the sample were streaked onto
Sorbitol MacConkey agar plates and incubated at 37  C for a maximum of All E. coli O157 isolates were subjected to susceptibility testing
24h. Following incubation, sorbitol negative (colorless) colonies were against eight antibiotics. Gentamycin, Kanamycin, Chloramphenicol,
identified by their color and further streaked onto sorbitol MacConkey Ciprofloxacin, Streptomycin, Tetracycline, Nalidixic acid, and Ampicillin
agar plates again to get a clear colorless typical E. coli O157 isolates. From disks were used to assess the susceptibility pattern of the isolates. The
the pure culture, isolates were sub-cultured onto nutrient agar for further method used for the susceptibility test was the Kirby-Bauer disk diffusion
preservation and transportation of isolates with glycerol stocks. on Muller-Hinton agar plates prepared according to the manufacturer's
recommendation. The isolated bacterial colonies from pure fresh culture

3
A. Assefa et al. Heliyon 5 (2019) e02996

Table 1. Susceptibility decision criteria for the Enterobacteriaceae (CLSI, 2015). Table 2. Prevalence of E. coli O157 in different sample types, Lakes and sampling
sites.
Antibiotic disk Disc concentration Zone of inhibition to the nearest mm
Variables Total Positives(n) p-value Crude OR 95%CI
Susceptible Intermediate Resistant sampled
Ampicillin 10 μgm 17 14–16 13 Sample type
CAF 30 μgm 18 13–17 12 Filleted fish swab 125 3 Reference
Ciprofloxacin 5 μgm 21 16–20 15 Filleted fish muscle (tissue) 89 2 0.942 0.93 0.15, 5.7
Gentamycin 10 μgm 15 13–14 12 Whole fish (skin) swab 125 1 0.3 0.33 0.03, 3.2
Kanamycin 30 μgm 18 13–17 13 Knife and cutting board swab 10 0 - * -
Nalidixic acid 30 μgm 19 14–18 13 Ready to eat fish 13 0 - * -
Streptomycin 10 μgm 15 12–14 11 Workers hand swab 24 0 - * -
Tetracycline 30 μgm 15 12–14 11 Container swab 24 0 - * -
CAF, chloramphenicol. Study Lake
Hayq 250 4 Reference
Tekeze 160 2 0.77 1.28 0.23, 7.09
were transferred to a test tube of 5 ml Trypton soya broth (TSB) and
Sampling Sites
incubated at 37  C for 6 h. The turbidity of the culture broth was adjusted
Landing site 293 5 Reference
using a sterile saline solution to obtain turbidity comparable to 0.5
Retail market 75 1 0.82 0.778 0.089, 6.76
McFarland turbidity standard. A sterile cotton swab was immersed in the
Restaurants 42 0 - * -
suspension and swabbed uniformly on the surface of Mueller-Hinton agar
plates. After the plates dried, antibiotic disks were placed and gently Total 410 6

pressed using sterile forceps and incubated at 37  C for 24 h. The *

Omitted values because of the absence of positive observations.
diameter of the inhibition zone formed around each disk was measured
using a digital caliper. The results were interpreted according to clinical nalidixic acid, and ampicillin. Antibiotics were selected based on their
laboratory standard protocols (CLSI, 2015) depicted in the following usage level and availability in the market. Based on the susceptibility test
Table 1. result, Ampicillin and Streptomycin performed poorly against the iso-
lates. On the other hand, Ciprofloxacin and Nalidixic acid were found to
2.6. Data management and analysis be effective in preventing the growth of most of the isolates. Tetracycline,
Gentamycin, and kanamycin also performed reasonably well.
All data generated in the field and laboratory were entered, coded, The maximum number of resistances recorded was only resistant to
and filtered in Microsoft Excel® version 2016 software. From the excel two drugs in five of the isolates. This indicates there were no multi-drug
sheet, data were further exported and analyzed using Stata 14 (Stata- resistant isolates. The overall resistance and resistance index pattern
Corp. Stata, Statistical Software: Release 14. College Station, TX) for were calculated and depicted in Table 3.
statistical handling purposes. Descriptive statistics (frequency tables and
graphs) were used to visualize the findings, while logistic regression and 4. Discussion
Fisher's exact test were used to make statistical inferences about the
findings. A p-value  0.05 was considered as significant. Escherichia coli O157: H7 is one of the most feared foodborne path-
ogens worldwide. It is a reportable organism in the USA and Europe. The
3. Results organism has been investigated for its occurrence in different foodstuffs
as well as cattle globally. In Ethiopia, there is a limitation in estimating
The overall prevalence of Escherichia coli O157: H7 in fish was found and identifying the occurrence of the organism in foods as well as cattle
to be 1.46% (6/410). Samples analyzed were filleted fish swab, filleted and human patients. However, one can argue that recently, there is a
fish muscle (tissue), whole fish swab, working knife and cutting board good trend of studying the organism in different foods of animal origin.
swab, ready to eat fish from restaurants, workers hand swab, and Reports in milk and meat estimated the overall prevalence of the or-
container swab). The occurrence of the organism was numerically higher ganism to be 4.9% (Assefa, 2019). The present study reported that the
in Lake Hayq than Tekeze Dam. Besides, it was also higher in filleted fish organisms’ occurrence in fish was 1.46% in the two study sites. This
than whole fish swabs. However, these differences were not found sta- prevalence is, however, lower than reported elsewhere from fish and
tistically significant. The organism was not isolated from ready to eat fish
sampled from restaurants, knife and cutting board sab, workers hand,
and container swabs. The prevalence of organisms was numerically Table 3. Antimicrobial resistance pattern of the isolates.
higher in samples taken from landing sites than local retail markets and Isolate Antibacterial Sample type Source Number of resisted MDR
restaurants. The prevalence of the organism in different sample types, code resistance profile (study Lake) antibiotics index
sampling location, and study Lakes is presented in Table 2. P1 S&AMP Whole fish Hayk 2 0.25 (2/
Because there were no enough positive observations in some vari- (skin) swab 8)
ables, the data were not suitable for chi-square analysis. Due to this issue, P2 S&AMP Filleted fish Hayq 2 0.25 (2/
Fisher's exact test analysis was performed. However, Fisher's exact sta- swab 8)
tistics do not show the direction of the association between the occur- P3 S&AMP Filleted fish Hayk 2 0.25 (2/
swab 8)
rence of the organism and risk factors included. To overcome this
problem, we used logistic regression analysis. However, due to a few P4 S&AMP Filleted fish Hayq 2 0.25 (2/
flesh 8)
positive observations, some of the variable levels were omitted during
P5 S&AMP Filleted fish Tekeze 2 0.25 (2/
analysis. A univariable logistic regression analysis was performed and the
swab 8)
crude odds ratio with P-value and 95% confidence interval was extracted
P6 AMP Filleted fish Tekeze 1 0.125
and displayed (Table 2). flesh (1/8)
Antibiotic susceptibility test was performed against antibiotic disks of
S¼Streptomycin, AMP ¼ Ampicillin MDR ¼ multi drug resistance.
gentamycin, kanamycin, CAF, ciprofloxacin, streptomycin, tetracycline,

4
A. Assefa et al. Heliyon 5 (2019) e02996

related species. A study conducted in the USA reported a prevalence of the need for further investigation of environments of landing sites that
6.95% in fish (Ribeiro et al., 2015), while in India from shrimp samples, pose a risk of contamination during processing.
the prevalence was found to be 5.8% (Surendraraj et al., 2010). In As E. coli is an indicator organism for the presence of contamination in
France, the organism was isolated from shellfish, with a prevalence of food samples, the existence of other pathogenic organisms in fish har-
4.16% (Gourmelon et al., 2006). This difference might be due to the level vested from those areas is imminent. Future researches may focus on
of sensitivity of tests used. detecting other pathogenic bacteria and their level of concentration in
The prevalence of the organism in this particular study was somehow fish. Besides, risk analysis of contracting infection due to E. coli O157 and
lower than other foods of animal origin in Ethiopia, as well. This dif- other pathogenic organisms from fish consumption should be analyzed
ference can be attributed to the fact that fish is a cold blood animal in based on data generated from every value chain.
which the organism cannot flourish in it. The source for the occurrence of The antibiotic sensitivity test result showed that essential drugs in the
the organism can be poor handling practice by fishers and retailers field of veterinary and human medicine performed reasonably well.
operating in the study sites. As noted during the study period, fishermen While drugs like Streptomycin and Ampicillin were poor in inhibiting the
process fish in soil, grass, and stones without any clean bedding material. growth of the isolates. Even though some illness conditions caused by this
This unhygienic handling practice can be one of the sources of contam- organism may not be treated by antibiotics at all, this situation is an
ination. In addition to this, fishermen do not prevent or reduce the ac- indication that Streptomycin and Ampicillin should not be used as a
tivity of flies that wonder in the fish during processing and marketing. A treatment option for infections caused by this organism. Recent studies
study showed that Escherichia coli O157 occurrence was 2.7 times higher showed that the organism demonstrated a high level of resistance in
in house flies (Musca domestica) than manure from dairy farms (Burrus Ethiopia as well as elsewhere for these drugs (Abdissa et al., 2017; Bedasa
et al., 2016), in which the authors recommended the diagnosis of flies et al., 2018; Sunde and Norstr€ om, 2005). It is also said that Streptomycin
that wander in restaurants can give a better result than foods sold. In light and Ampicillin are the two most frequently co-transferred resistance
of this, the occurrence of Escherichia coli O157: H7 in fish can also be phenotypes among sulfonamide-resistant E. coli isolates recovered
attributed to the free activity of flies in the study locations. (Sunde and Norstr€ om, 2005). Chloramphenicol, one of the most
The occurrence of Escherichia coli O157: H7 was numerically higher in frequently and widely used drug, performed reasonably well in inhibiting
Lake Hayq than Tekeze dam harvested fish. This difference can be due to the growth of 4 the isolates. However, one isolate was found to be an
1) the sample size is relatively larger in Lake Hayq, 2) higher agricultural intermediately resistant to this drug.
practices around Lake Hayq were noted. Farmers that live near Hayq Even though there were no resistant isolates for at least three drugs,
water their cattle in the Lake, graze around the Lake where fish is gutted yet the multidrug resistance index calculated showed a relatively higher
and filleted, they also use cattle manure to crop papaya, sugarcane, ba- index. From the total six isolates, five of them were resistant to two drugs
nana, onion and many more crops around the Lake. The remaining of the (Ampicillin and Streptomycin), while one isolate was resistant to ampi-
manure drain to the landing site where fish is filleted and arranged for cillin only. Hence the MDR index of the isolates was found between 0.125
sale, 3) Lake Hayq is relatively humid than Tekeze (Mulugeta et al., and 0.25. This value is somehow higher than the accepted international
2017) which makes it a favorable environment for the organism to thrive standard for a multi-drug resistance index of 0.2. This can be due to
and for increased activities of flies than the hot climate Tekeze area. widespread irrational drug use in both animals and humans in the
The prevalence of the organism was also found numerically higher in country. Irrational drug use can result in resistant gens to develop and
filleted fish (swab and flesh) than whole fish samples. This may indicate further transfer to susceptible organisms from the environment (Von
that the processing environment (filleting and gutting) can be a source of Baum and Marre, 2005). This situation may need attention sooner than
contamination. A whole fish harvested from the Lakes may have not later before losing effective drugs regarding cost and efficiency. Even
enough time to be contaminated from the environment than filleted fish. though it is out of the scope of this paper, it is wise that responsible
During the process of filleting, almost all parts of the body come in bodies like drug authority and control agency (DACA) should identify
contact with the ground. Consequently, causing a higher contamination actors involved in irresponsible drug use; develop and implement strict
rate than whole fish indicating unhygienic processing may expose fish to control measures to the adhesion of rational drug use at the national
this and other pathogenic organisms. level.
Regarding the sampling sites, the organism's prevalence was numer-
ically higher at landing sites than nearby retail markets and restaurants. 5. Conclusion
In Tekeze there were no local markets as well as restaurants in the nearby
district (Abergele) that serve fish. Fish harvested in that Dam is directly This study is the first of its kind in Ethiopia in investigating
sold in landing sites for whole sellers who store in the freezer and load it Escherichia coli O157: H7 from fish. Handling practices by fishermen
to Addis Ababa and Mekelle cities. Due to this reason, samples from during filleting and gutting might be the cause of contamination.
restaurants and retail markets were taken only from Hayq city. Hence, Ready to eat fish sampled from restaurants were found free of the
the sample size from the landing site, retail market, and restaurants may organism indicating that it may not reach end-users provided that
not be proportional. This situation further can lead us to the conclusion proper cooking is in place. Since this study has a limitation of
that the higher occurrence of the organism in landing sites can be due to a including enough sample size from restaurants, this argument may
higher number of fish samples taken from landing sites. The other need further investigation of fish serving restaurants with a repre-
conclusion can be, in landing sites, fish filleted and mishandled exposed sentative sample size. Drugs like ciprofloxacin, gentamycin, tetracy-
to dust, dung, and flies. However, fishermen wash their harvest after cline, and nalidixic acid were effective in inhibiting the growth of the
finishing the filleting process which may lower the chance of recovery of isolates while streptomycin and ampicillin performed poorly. In light
the organism further in the value chain. of this, raw and undercooked fish consumptions in Ethiopia may result
Though we did not sample ready enough to eat fish from restaurants, in contracting infections. Since this study identified a single bacterium,
the occurrence level of the organism was zero out of thirteen samples other critical foodborne pathogens associated with fish should be
examined. This is somehow a relief that the organism may not reach to routinely examined. Besides, future studies shall consider taking
end-users. However, this argument may need further detailed investi- samples from the environments, the lake water as well as livestock
gation with a representative sample size in fish serving restaurants to grazing around that areas, particularly cattle as they are the main
have a sound conclusion. In addition to restaurants, the organism was not carriers of E. coli strains, running water that can bring cattle feces to
also isolated from other working environment samples. Containers, knife, the lake and then looking at the genetic relatedness of the isolates from
and workers' hand swabs were free from the organism. This suggests that these different sources shall be the focus of research.

5
A. Assefa et al. Heliyon 5 (2019) e02996

Declarations Desta, H., Donata, C., Gabriella, S., Martino, P., 2016. Bacteriological quality of milk in
raw bovine bulk milk in the selected milk collection centers: smallholder dairy
processing Ethiopia. J. Vet. Sci. Anim. Husb. 4, 1–5.
Author contribution statement Earley, B., Leonard, N., 2006. The characterisation of E . coli O157 : H7 isolates from
cattle faeces and feedlot environment using PFGE. Vet. Microbiol. J. 114, 331–336.
Fikru Regassa: Conceived and designed the experiments; Analyzed Elsaidy, N., Abouelenien, F., Kirrella, G.A.K., 2015. Impact of using raw or fermented
manure as fish feed on microbial quality of water and fish. Egypt. J. Aquat. Res. 41.
and interpreted the data; Contributed reagents, materials, analysis tools FAO, 2016. Aquacultue Summary: the State of World Fisheries and Aquaculture Report.
or data. FAO, pp. 1–8, 2016.
Dinka Ayana, Kebede Amenu: Conceived and designed the experi- Garcia-Rodríguez, F.G., De La Cruz-Aguero, J., 2011. Review on post-harvest losses in
artisanal fisheries of some African countries. Fish. Aquat. Sci. 6, 186–193.
ments; Contributed reagents, materials, analysis tools or data. Gourmelon, M., Montet, M.P., Lozach, S., Le Mennec, C., Pommepuy, M., Beutin, L.,
Fufa Abunna: Conceived and designed the experiments. Vernozy-Rozand, C., 2006. First isolation of Shiga toxin 1d producing Escherichia coli
Ayalew Assefa: Conceived and designed the experiments; Performed variant strains in shellfish from coastal areas in France. J. Appl. Microbiol. 100,
85–97.
the experiments; Analyzed and interpreted the data; Contributed re- Haileselassie, M., Taddele, H., Adhana, K., Kalayou, S., 2013. Food safety knowledge and
agents, materials, analysis tools or data; Wrote the paper. practices of abattoir and butchery shops and the microbial profile of meat in Mekelle
City, Ethiopia. Asian Pac. J. Trop. Biomed. 3, 407–412.
Hiko, A., Asrat, D., Zewde, G., 2008. Occurrence of Escherichia coli O157:H7 in retail raw
Funding statement meat products in Ethiopia. J. Infect. Dev. Ctries. 2, 389–393.
Hossain, M.M., Hassan, M.N., Sayem, S.M., Polanco, J.F., 2015. Assessment of post
This work was supported by Addis Ababa University, research harvest loss of wet fish: a novel approach based on sensory indicator assessment.
SAARC J. Agric. 13, 75–89.
directorate and Amhara Agricultural Research Institute.
ISO6887-3:2017, 2017. Microbiology of the Food Chain — Preparation of Test Samples,
Initial Suspension and Decimal Dilutions for Microbiological Examination — Part 3:
Competing interest statement Specific Rules for the Preparation of Fish and Fishery Products.
Lopez-Campos, G., Martinez-Suarez, J.V., Aguado-Urda, M., Alonso, V.L., 2012.
Detection, identifi cation, and analysis of foodborne pathogens. In: Microarray
The authors declare no conflict of interest. Detection and Characterization of Bacterial Foodborne Pathogens. Springer Briefs in
Food, Health, and Nutrition, pp. 13–33.
Additional information Mulugeta, M., Tolossa, D., Abebe, G., 2017. Description of long-term climate data in
Eastern and Southeastern Ethiopia. Data Br. 12, 26–36.
Ribeiro, L.F., Barbosa, M.M.C., Pinto, F.D.R., Guariz, C.S.L., Maluta, R.P., Rossi, J.R.,
No additional information is available for this paper. 2015. Shiga toxigenic and enteropathogenic Escherichia coli in water and fish from
pay-to-fish ponds. Lett. Appl. Microbiol. 62, 216–220.
Sanath Kumar, H., Otta, S.K., Karunasagar, I., Karunasagar, I., 2001. Detection of Shiga-
References toxigenic Escherichia coli (STEC) in fresh seafood and meat marketed in Mangalore,
India by PCR. Lett. Appl. Microbiol. 33, 334–338.
Abdissa, R., Haile, W., Fite, A., Beyi, A., Agga, G., Edao, B., Tadesse, F., Korsa, M., Seid, Z., 2016. Fish Diversity, Abundance, Socioeconomic Importance and the Status of
Beyene, T., De Zutter, L., Cox, E., Goddeeris, B., 2017. Prevalence of Escherichia coli the Fisheries of Lake Hayq, South Wollo, Ethiopia. Addis Ababa university.
O157:H7 in beef cattle at slaughter and beef carcasses at retail shops in Ethiopia. Sunde, M., Norstr€ om, M., 2005. The genetic background for streptomycin resistance in
BMC Infect. Dis. 17, 277. Escherichia coli influences the distribution of MICs. J. Antimicrob. Chemother. 56,
Abebe, M., Hailelule, A., Abrha, B., Nigus, A., Birhanu, M., Adane, H., Genene, T., 87–90.
Daniel, H., Getachew, G., Merga, G., Haftay, A., 2014. Antibiogram of Escherichia coli Surendraraj, A., Thampuran, N., Joseph, T., 2010. Molecular screening, isolation, and
strains isolated from food of bovine origin in selected Woredas of Tigray , Ethiopia. characterization of enterohemorrhagic Escherichia coli O157:H7 from retail shrimp.
J. Bacteriol. Res. 6, 17–22. J. Food Prot. 73, 97–103.
Agüeria, D.A., Terni, C., Baldovino, V.M., Civit, D., 2018. Food safety knowledge, Tassew, H., Abdissa, A., Beyene, G., Gebre-Selassie, S., 2010. Microbial flora and food
practices and attitudes of fishery workers in Mar del Plata, Argentina. Food Control borne pathogens on minced meat and their susceptibility to antimicrobial agents.
91, 5–11. Ethiop. J. Health Sci. 20, 137–143.
Albuquerque, R., 2013. Escherichia coli in seafood: a brief overview. Adv. Biosci. Taye, M., 2013. Study on carcass contaminating Escherichia coli in apparently healthy
Biotechnol. 4, 450–454. slaughtered cattle in Haramaya university slaughter house with special emphasis on
Assefa, A., 2019. Prevalence of Escherichia coli O157:H7 in foods of animal origin in Escherichia coli O157:H7, Ethiopia. J. Vet. Sci. Technol. 4, 4–6.
Ethiopia: a meta-analysis. Cogent Food Agric. 1–10, 0. Teame, T., Natarajan, P., Tesfay, Z., 2016. Assessment of fishery activities for enhanced
Bedasa, S., Shiferaw, D., Abraha, A., Moges, T., 2018. Occurrence and antimicrobial management and improved fish production in Tekeze reservoir , Ethiopia. Int. J.
susceptibility profile of Escherichia coli O157:H7 from food of animal origin in Fauna Biol. Stud. 3, 105–113.
Bishoftu town, Central Ethiopia. Int. J. Food Contam. 5, 2. Tesfaye, G., Wolff, M., 2014. The state of inland fisheries in Ethiopia: a synopsis with
Bevilacqua, A., Corbo, M.R., Sinigaglia, M., 2017. The Microbiological Quality of Food. updated estimates of potential yield. Integr. Med. Res. 14, 200–219.
Woodhead Publishing, Amsterdam. Thrusfield, M., 2007. Veterinary Epidemiology, 3rd. Ed. Blackwell Science, Oxford.
Burrus, R.G., Hogsette, J.A., Kaufman, P.E., Maruniak, J.E., Simonne, A.H., Mai, V., 2016. Von Baum, H., Marre, R., 2005. Antimicrobial resistance of Escherichia coli and
Prevalence of Escherichia coli O157:H7 from house flies (Diptera: Muscidae) and therapeutic implications. Int. J. Med. Microbiol. 295, 503–511.
dairy samples in North central Florida 1. J. Med. Entomol. 1–9, 0. Zhao, C., Zhang, Y., Yang, S., Xiang, H., Sun, Y., Yang, Z., Yu, Q., Lim, R.P., 2018.
CLSI, 2015. In: Clinical and Laboratory Standards Institute (Ed.), Performance Standards Quantifying effects of hydrological and water quality disturbances on fish with food-
for Antimicrobial Disk and Dilution Susceptibility Tests for Bacteria Isolated from web modeling. J. Hydrol 560, 1–10.
Animals, CLSI Docum. Clinical and Laboratory Standards Institute, Wayne, PA.
Cobbaut, K., Houf, K., Buvens, G., Habib, I., Zutter, L. De, 2009. Occurrence of non-
sorbitol fermenting , verocytotoxin-lacking Escherichia coli O157 on cattle farms.
Vet. Microbiol. J. 138, 174–178.