Introduction
Diarrheal diseases are the leading cause of childhood illness and death in developing countries (1). Enteric pathogens including rotavirus, norovirus, Campylobacter, Salmonella, E. coli and Shigella species are known causative agents (2, 3). Campylobacter species are the leading zoonotic bacterial causes of human gastroenteritis and foodborne disease globally (1, 2, 4, 5). The genus Campylobacter, comprises of 31 species and 13 subspecies (6); of these, thermophilic Campylobacter, especially C. jejuni and C. coli, are the primary causes of human bacterial gastroenteritis worldwide (2, 7).
Epidemiologically, human campylobacteriosis differs across high and low-income countries. In high-income countries, symptomatic infection occurs in all age groups, whereas in low-income countries, clinical disease mostly affects children under 5 years of age, with adults rarely suffering from the disease, predominately having asymptomatic excretion (8, 9). Most Campylobacter species have been isolated from animals, and direct and indirect contact with animals is a known risk factor for Campylobacter infection, especially in children (10–12). Animals can serve both as reservoirs, and as potential sources for the contamination of food, water, and the environment (4, 13–15). Contamination of different water sources with Campylobacter species presents an opportunity for transmission of environment-adapted genotypes to livestock and humans (7, 13, 14, 16).
Most cases of campylobacteriosis are self-limiting, and present with enteritis, abdominal cramps, fever, nausea, and vomiting (3, 9). However, some cases have been linked to more serious complications including Guillain–Barré syndrome, Reiter's syndrome, bacteremia, and abortion (9). In severe cases of human campylobacteriosis, macrolides, and fluoroquinolones (particularly erythromycin and ciprofloxacin), and intravenous aminoglycosides are the first line antimicrobials of choice (14, 17). There is increasing antimicrobial resistance amongst Campylobacter isolates from various sources (18–20). This situation is more common in developing countries, where there is widespread and largely uncontrolled use of antimicrobials (11, 12, 21).
Studies on the prevalence of Campylobacter in Ethiopia have identified a range of 13.8–50.0% in humans, and 10.6–56.5% in animals (12, 13, 22–27). In addition, studies conducted on the antimicrobial susceptibility profiles of Campylobacter species, in both humans and animals in Ethiopia, have identified increased multidrug resistance of Campylobacter species (4, 25, 28). Most of these studies were conducted in only humans or animals, and utilized only a culture method approach for Campylobacter species identification and characterization. The role of different environments in Ethiopia, including water, as potential risk factors and sources for transmission of Campylobacter infection in humans and animals is far less known. In addition, there is limited literature regarding risk factors associated to human infection with Campylobacter species in Ethiopia. The identification and characterization of Campylobacter species in humans, animals, and water sources, is important in efforts to reduce the risk of human exposure to Campylobacter. The objectives of this study were the identification and characterization of Campylobacter spp. in humans, animals and water sources in livestock owning households of peri-urban Addis Ababa, Ethiopia; and the characterization of antimicrobial resistant isolates.
Materials and Methods
Study Design and Location
A cross-sectional study was conducted between December 2018 and January 2020 in Addis Ababa, Ethiopia. Addis Ababa has 10 sub-cities, of which five are considered as peri-urban areas. Akaki Kality sub-city was selected as it is characterized by a high level of livestock production, and studies have documented a high prevalence of zoonotic enteric infections (29–31). Within Akaki Kality sub-city, two woredas (Woreda 2 and 3) were purposively selected.
Study Population
Study participants were selected from available lists of livestock-owning households in both woredas using simple random selection. Randomization of selected household lists and households was conducted using random numbers generated in a software program (Microsoft Excel 2019, Microsoft Cooperation, USA). Sample sizes were calculated for two independent populations based on an expected Campylobacter species prevalence of 50%, 5% margin of error, and 95% confidence level, using the total number of available livestock-owning households (227 households) in Woreda 2 and 3. The calculated sample size was 143 households and 372 animals. Households were included in the study if they owned at least one free-roaming chicken, or at least one ruminant animal. Humans and animals that were on antimicrobial treatment, or who were treated during the previous 2 weeks at the time of contact, were excluded from the study.
Of the 227 households contacted, 99 consented to participate. Households refused to participate for the following reasons: recently lost, or sold livestock (n = 94), no reason provided (n = 21), and refused to provide human or animal fecal samples (n = 13). Therefore, 99 households participated, providing 99 household pooled human fecal samples, and 248 household pooled livestock fecal samples. The sample size for water samples was conveniently determined based on the available water sources that was in use for both household daily activities, and drinking water for humans and animals in the study location (14).
Questionnaire data on potential risk factors associated with human infection with Campylobacter species were collected. Both closed and open-ended questions were utilized and translated into local languages (Afaan Oromoo and Amharic). The questionnaire was pretested on 20 households from a similar Woreda (8) in Yeka sub-city.
Sample Collection
A total of 347 household pooled fecal samples were collected from households: humans (n = 99), cattle (n = 135), sheep (n = 30), goats (n = 14), and poultry (n = 69). Pooled human stool samples were collected only from individuals who were most closely associated with the management of the livestock. Water samples (n = 172) were collected from the 99 households: surface water (n = 9), municipal pipe water (n = 84), ground water (n = 16), and stored water (n = 63).
A sterile cotton swab moistened with nutrient broth was used to transfer ~7–10 g of the pooled fecal samples into a 15 ml screw-capped falcon tube containing Cary-Blair Transport Medium (Oxoid). Fecal samples from individual animals were collected either directly from rectum or floor of the livestock housing immediately after an animal defecated. Samples were transported to the Microbiology Laboratory of Akililu Lemma Institute of Pathobiology (ALIPB), Addis Ababa University (AAU), and processed within 4–6 h of collection.
Isolation and Identification of <italic>Campylobacter</italic> Species
A selective enrichment was initially performed on all samples as previously described (32). Water samples were filtered through a 0.45 μm sterile nitrocellulose filter (Pall Corporation, Ann Arbor, MI). Membrane filtered water and homogenized human and animal fecal specimens were inoculated in 30 ml of Bolton enrichment broth containing Campylobacter growth supplement (HiMedia Laboratories; Mumbai, India), with 5% (v/v) defibrinated cattle blood. Tubes were then incubated for 48 h under microaerophilic condition (CO2, 10%; O2, 5%; N2, 85%) using Campy Gene kits (Thermo Scientific, Waltham, MA, USA) as described by Szczepanska et al. (14).
After 48 h of incubation, a loopful of culture from each sample type was spread onto plates containing modified charcoal cefoperazone deoxycholate agar (mCCDA, CM0739-Preston) and Campylobacter selective supplements (SR0174), and incubated under microaerophilic conditions as described earlier for 48–72 h at 42°C. After 48–72 h of incubation, all presumptive Campylobacter colonies were selected from each plate and checked for motility, oxidase, catalase and Gram character as previously described (7). All suspected colonies were subsequently sub-cultured onto Columbia blood agar (Difco, USA) with 5% (v/v) defibrinated cattle blood and incubated under microaerophilic condition at 42°C for 48 h. Isolates with typical colony characteristics of Campylobacter were then subjected to PCR analysis for confirmation and identification of Campylobacter species. Reference strains of Campylobacter jejuni and coli were obtained from the Ethiopian Public Health Institute and used as positive controls.
Identification of <italic>Campylobacter</italic> by PCR
Genomic DNA was extracted by boiling fresh Campylobacter cultures grown on Columbia blood agar as previously described (33). A loopful of bacterial growth from plates were suspended in 100 μl of sterilized RNase/DNase- free water, boiled at 95°C for 15 min and cooled at 4°C until used. Genus level confirmation and identification of Campylobacter was conducted using multiplex-PCR (mPCR) targeting genus level and species-specific regions of the 16S rRNA, j0414, glyA, cstA, and ask gene sequences, as described previously and in Supplementary Table 1 (34). Campylobacter identification was conducted and isolates that were positive for the genus-specific PCR but negative for the C. lari, C. fetus, C. coli, and C. jejuni- using mPCR assay were designated as unidentified thermophilic Campylobacter species. In this study, a household was considered positive for Campylobacter, if at least one isolate was confirmed by PCR from the pooled samples of any of the animal, human, or water samples obtained.
Antimicrobial Susceptibility of <italic>Campylobacter</italic> Isolates
All PCR confirmed isolates were subjected to antimicrobial susceptibility testing using the standard Kirby-Bauer disc diffusion assay on Mueller-Hinton agar (Hi Media Laboratories; Mumbai, India) according to Clinical Laboratory Standards Institute (CLSI) guidelines (35). A panel of 11 antimicrobials were used from BD BBL Sensi-Discs: ciprofloxacin (CIP) (5 μg), nalidixic acid (NA) (30 μg), erythromycin (E) (15 μg), azithromycin (AZM) (15 μg), tetracycline (Te) (30 μg), gentamicin (GM) (10 μg), ampicillin (AM) (10 μg), amoxicillin-clavulanic acid (AMC) (10 μg), chloramphenicol (C) (30 μg), and trimethoprim-sulfamethoxazole (SXT) (25 μg), and amikacin (AN) (30 μg). Isolates were defined as multidrug-resistant (MDR) when resistant to three or more classes of antimicrobials (36).
Data Analysis and Ethical Approval
Data were analyzed using SPSS v25 (SPSS Inc, Chicago, Illinois, USA). Descriptive statistics were calculated for Campylobacter species prevalence, antimicrobial susceptibility and household demographic data. Chi-squared and Fisher's tests were used as appropriate to compare differences in the prevalence and antimicrobial resistance profiles between Campylobacter species isolates, and between strains isolated from different sources. Potential risk factors associated with the occurrence of Campylobacter in human samples were analyzed using univariable logistic regression. The outcome measure used was a binary variable reflecting whether a household was PCR positive for Campylobacter. Variables were checked for collinearity, and a backward-stepwise process was used, with covariates remaining in the model if they were statistically significant (p < 0.05). All variables that showed an association with the outcome variable on univariable analysis (p < 0.2) were considered in the final multivariable logistic regression analysis. Ethical approval was obtained (DERC/18/19/10-A) from the Research Ethics Review Committee (RERC) of the Department of Microbiology, Immunology and Parasitology (DMIP), College of Health Sciences, Addis Ababa University. Permission was obtained from the Addis Ababa Bureau of Agriculture and Livestock and Addis Abba Bureau of Health. Written informed consent was obtained prior to enrolment from the head of each household for participation in this study.
Discussion
Of the 99 households in this study, 42.4% were positive for Campylobacter species in at least one of the household samples tested. The prevalence of Campylobacter species in humans in our study, was similar to a number of previous studies in Ethiopia (20, 38), Poland (14), and Tanzania (39). In contrast, studies in Ethiopia by Terefe et al. (22), Tafa et al. (23), and Lengerh et al. (40) reported higher prevalences than our study. A higher prevalence was also reported by Chuma et al. (41) in Tanzania, and Schiaffino et al. (36) in Peru, whereas Rawat et al. (37) and Meistere et al. (42) reported a lower prevalence in Latvia and India. The variation in prevalence across the studies might be due to differences in study methodologies, periodical and environmental variations, and differences in the study groups (14, 42).
Campylobacter species prevalence observed in cattle in this study is consistent with a previous Ethiopian study (43). However, it is a higher prevalence than reported by Hagos et al. (4) and Kassa et al. (25) in Ethiopia, and a lower prevalence than reported by Abamecha et al. (44) in Ethiopia. Comparable results were also reported from the Republic of Korea (45) and Latvia (42). Prevalence in our study was higher than those reported from Iran (46), and Ghana, but lower than reports from Republic of Korea (47), and the USA (48). The prevalence of Campylobacter species observed in poultry in this study is consistent with the report from Latvia (42), but lower than previous studies in Ethiopia (12, 25), Tanzania (41), Kenya (49), and Cambodia (15). However, the prevalence in our study is higher than the reports by Pires et al. (50) from California, USA, and Rawat et al. (37) from India.
The variation in prevalence of Campylobacter species in animals across these studies might be explained by the differences in study methodology and duration, seasonality, animal management system, sanitary practices, and agro-ecological variations (15, 41). Certain studies, like Hagos et al. (4), reported a prevalence from meat products rather than live animals. Our study was a community level study, utilizing a pooled sampling approach, where all sampling units were asymptomatic at the time of sample collection. In contrast, many previous studies conducted in Ethiopia were based on symptomatic human subjects, and/or farm-level animal studies that potentially resulted in the observed variation in prevalence.
Our study utilized both a culture and molecular approach for the isolation and identification of Campylobacter species. The majority of previous studies in Ethiopia used only the culture and biochemical test approach for the isolation and identification of Campylobacter. This latter approach has the potential to be less sensitive and specific, and could be a reason for the observed differences in prevalence (51–53). Nevertheless, the higher prevalence of zoonotic thermophilic Campylobacter species observed in livestock in our study is of considerable concern, as poultry and other livestock species move freely around the household, contaminating the environment, and are a source of infection for humans, especially children.
Of all the sample types in our study, the prevalence of Campylobacter species was highest in surface water samples (22.5%). The overall prevalence of Campylobacter species observed in all water types in our study, is consistent with a study from Turkey, but lower than studies from South Africa (7) and Poland (14). The variation across studies might be attributed to differences in the sources of water, water collection approach, season, geographical location, and method of isolation. In addition, our study area is impacted by both agricultural operations (cattle, poultry, and vegetable production), and urban wastewater treatment, which may influence the prevalence of Campylobacter and other microorganisms (31, 54). Although the contribution of water to the burden of Campylobacter infection in humans might be unknown (55), the results of our study provide information regarding the potential transmission of Campylobacter species between animals, humans and water sources. More importantly, our study revealed a high prevalence of thermophilic Campylobacter species in municipal tap water, indicating a potential risk for human infection. However, this result should be interpreted with caution, as no information was obtained on whether the source water was treated. The significantly higher isolation of Campylobacter from households using stored water, is consistent with previous studies (7, 56). Studies have shown that untreated stored water is a significant source of Campylobacter infections and outbreaks (16, 55). Risk factors including the source of water and type of storage container have been linked to the poor microbial quality of stored household water (56).
In our study, C. jejuni was the most frequently isolated species from all sample types, except for cattle and poultry isolates. This is consistent with previous studies in Ethiopia (22, 25) and other countries, including South Africa and California, USA (7, 50). C. jejuni is more prevalent, and has a longer viability in the environment compared to other thermophilic Campylobacter species, therefore increasing its chance of recovery (33, 50). This difference could also occur due to variations in the mechanism of pathogenesis and elimination amongst the different thermophilic Campylobacter species within the host cells (14). Differences in the isolation of Campylobacter species might also be related to their actual compositional variations in local environments. C. lari was not recovered in our study, consistent with previous studies in Ethiopia (22, 28, 33, 50). As demonstrated in other studies (7, 12, 39), our study also highlighted that Campylobacter species isolated from human feces also co-occurred in other samples (predominately cattle, poultry, and water) from the same household, suggesting that cattle, water and poultry are potentially the main sources of human Campylobacter infection. It is also possible, that humans could have infected the cattle, poultry and stored water.
The association of indoor and outdoor manure collecting with increased odds of human Campylobacter species positivity is consistent with previous studies (15, 47). Campylobacter can survive at variable rates in stored manure, and even in composted manure [reviewed in (50, 57)]. Consequently, human exposure to Campylobacter species may occur when drinking or ingesting of contaminated products (47). The negative association of taking any specific action to protect oneself while cleaning an animal house, or washing hands with soap before and after cooking, with Campylobacter species positivity is similar to a previous study (15). Poor hygiene and sanitation are associated with increased odds of multiple adverse health outcomes (58). However, our study indicated that there was no significant association between the consumption of raw or under cooked meat, and owning different animal species with human Campylobacter species positivity.
Nearly all isolates in our study were resistant to one or more antimicrobials. Most isolates were resistant to amikacin (79.1%), followed by amoxicillin-clavulanic acid (70.1%), whereas a lower resistance was observed to chloramphenicol (19.4%), followed by gentamicin (31.3%). Similar results have been reported previously in Ethiopia (59) and Iran (60). The level of resistance to nalidixic acid (64.2%) and ciprofloxacin (62.7%) observed in our study are consistent with other studies in Ethiopia (44), Poland (61), and Kenya (49). Resistance to erythromycin (61.2%) and azithromycin (59.7%) observed in this study, is consistent with results found in other previous studies (33, 44, 60), but higher than other studies (14, 23, 40). In our study, resistance observed to tetracycline (67.2%), ampicillin (64.2%), and sulfamethoxazole-trimethoprim (41.8%) is consistent with previous studies (7, 40), but lower resistance has been previously reported (25, 37).
In our study, C. coli was highly resistant to nalidixic acid, amoxicillin-clavulanic acid, azithromycin, and ampicillin, compared to other species. Whereas, C. jejuni showed relatively higher resistance to erythromycin, ciprofloxacin and amikacin than other species. These differences could be due to greater use of these drugs in humans, and that humans are commonly infected with these species of Campylobacter (62). This is consistent with studies from Ethiopia (13), South Africa (7, 63), Tunisia (64), but differs from other studies in Ethiopia (59), Egypt (65), and England (66).
The high number of multidrug resistant Campylobacter isolates observed in our study is consistent with previous studies (7, 44). However, lower antimicrobial resistance has been previously reported in Ethiopia (25), Iran (46), and Tanzania (33). One potential hypothesis for the high resistance exhibited by C. coli, compared to other Campylobacter species, may be that C. coli strains can acquire resistance genes horizontally more effectively than other species, and chromosomally encoded target genes could mutate faster in C. coli (67, 68). However, we need to be mindful about our interpretation, given the low number of C. coli isolates in our study.
The higher antimicrobial resistance observed in our study may potentially be due to the overuse, and inappropriate use, of antimicrobial agents in both human and animals. To our knowledge, oxytetracycline and penicillin (alone or in combination with streptomycin) are used widely in livestock and poultry in Ethiopia. This use contributes to increased selection of resistant Campylobacter species (69, 70). Extensive use of antimicrobials in the animal industry for prophylaxis, and growth promotion, has been associated with increased resistance (14, 65, 71). Due to the close human and animal interaction in our study area, there is the potential for antimicrobial resistant strains, or resistant genetic markers originating from humans, to be transmitted to animals. The differences in the isolation of Campylobacter species across the different sources, may make it challenging to compare the levels of resistance between the sources. Nonetheless, observed resistance discrepancies in our study, compared to previous studies, could be due to differences in exposure rates of the bacteria to the different antimicrobials. The inevitable human-animal-environment interaction in our study, along with inappropriate use of antimicrobials in both animals and humans in Ethiopia, might potentially lead to increased selection pressures for resistant strains.
The challenges presented by infectious diseases and antimicrobial resistance are multifaceted, and it is critically important to address these issues using a One Health approach (72, 73). One Health interventions advocate close intersectoral cooperation, interdisciplinary expertise, and the involvement, and empowerment of multiple stakeholders (74, 75). Our study demonstrated that C. jejuni, C. coli, C. fetus, and unidentified Campylobacter species were prevalent in the study area. The relatively high proportion of Campylobacter species in both livestock and water samples are potential risks for human Campylobacter infection. The high prevalence of Campylobacter species isolated from various water sources, highlights the need for further work to identify for how long Campylobacter species persist in these water sources, and to ascertain the transmission of Campylobacter within the environment. Multi-drug resistant zoonotic Campylobacter species were prevalent in animals, humans and environment in the peri-urban livestock owning households of Addis Ababa, Ethiopia. Recognizing the significant implications of antimicrobial resistance, it is important for Ethiopia to develop and implement a national plan to advance the rational use of antimicrobials utilizing a One Health approach. A One Health approach is recommended to further investigate Campylobacter species infections, and other zoonotic infectious agents, in the livestock owning populations in Ethiopia, where there is close interaction between humans, animals and the environment.