Campylobacter occurrence and antimicrobial resistance profile in under five-year-old diarrheal children, backyard farm animals, and companion pets

Campylobacteriosis disproportionately affects children under five in low-income countries. However, epidemiological and antimicrobial resistance (AMR) information at the children-animal interface is lacking. We hypothesized that Campylobacter is a major cause of enteritis in children in Ethiopia, and contact with animals is a potential source of transmission. The objective of the study was to determine Campylobacter occurrence and its AMR in children under five with diarrhea, backyard farm animals, and companion pets. Stool from 303 children and feces from 711 animals were sampled. Campylobacter was isolated through membrane filtration on modified charcoal cefoperazone deoxycholate agar plates under microaerobic incubation, and the technique showed to be feasible for use in regions lacking organized laboratories. Typical isolates were characterized with MALDI-TOF MS and multiplex PCR. Of 303 children, 20% (n = 59) were infected, with a higher proportion in the 6 to 11-month age group. Campylobacter occurred in 64% (n = 14) of dogs and 44% (n = 112) of poultry. Campylobacter jejuni was present in both a child and animal species in 15% (n = 23) of 149 households positive for Campylobacter. MICs using the gradient strip diffusion test of 128 isolates displayed resistance rates of 20% to ciprofloxacin and 11% to doxycycline. MICs of ciprofloxacin and doxycycline varied between C. coli and C. jejuni, with higher resistance in C. coli and poultry isolates. Campylobacter infection in children and its prevalent excretion from backyard poultry and dogs is a understudied concern. The co-occurrence of C. jejuni in animals and children suggest household-level transmission As resistance to ciprofloxacin and doxycycline was observed, therapy of severe campylobacteriosis should consider susceptibility testing. Findings from this study can support evidence-based diagnosis, antimicrobial treatment, and further investigations on the spread of AMR mechanisms for informed One Health intervention.


Introduction
Gastrointestinal diseases cause around 525,000 deaths yearly in children under five [1] and Campylobacter jejuni and Campylobacter coli have been ranked as one of the four principal microorganisms causing gastroenteritis worldwide [2,3].Campylobacteriosis occurs in all age groups in high-income countries, but it is widespread and disproportionately affects children under five in low-income countries [4].A pooled prevalence of 10% was reported in children under five, and 9% in a general population in Ethiopia [5,6].In children under five, it is strongly linked to severe morbidity, environmental enteric dysfunction, malnutrition, failure to thrive, neurological disorders and death [7].Children afflicted with campylobacteriosis experience fever, vomiting, abdominal cramps, watery to severe inflammatory diarrhea and dehydration [8].Furthermore, post-infection complications of campylobacteriosis include arthritis, Guillain-Barre ´syndrome resulting in respiratory and neurological dysfunction, and irritable bowel syndrome [9].
Ciprofloxacin is one of the first choice treatment options in severe cases of bacterial enteritis [10].Currently, children are however mostly treated without prior identification of the specific causative organism and antimicrobial susceptibility profiling in Ethiopia [11].The use of inappropriate antimicrobials for empirical treatment might contribute to the selection of resistant strains and AMR.Antimicrobial resistance can result from a genetic mutation within the organism or acquiring resistance genes from other organisms following selection pressures in humans and animals [12].
Campylobacter can colonize, persist, and exert different pathogenicity in a variety of hosts, and is considered the main cause of foodborne bacterial zoonotic infections [13].Poultry, pigs, and cattle are important reservoirs for the thermophilic Campylobacter species, C. jejuni and C. coli [14,15].Direct contact with feces of poultry, ruminants, and pets, as well as consumption of contaminated food of animal origin and drinking water, are considered the main transmission routes [16,17].Due to the broad host range and uncontrolled uses of antimicrobials in human and veterinary medicine, and agriculture, the emergence and dissemination of AMR in Campylobacter is both a global health challenge and threat to One Health [18][19][20].
Campylobacter is a Gram-negative curved or spiral, motile, microaerobic and fastidious bacterium, and not easy to cultivate in sub-optimal laboratory conditions, as in some parts of Ethiopia [2,3].As a result, insufficient attention is given to the epidemiological burden of disease and to optimal isolation and correct identification of Campylobacter from children with diarrhea.Because of variability in isolation procedures with respect to the choice of culture media and systems of microaerobic condition, temperature, and duration of incubation, results from previous studies are inconsistent.Aside from two recent studies, Terefe et al. [7] and Chala et al. [20], which applied molecular methods to detect Campylobacter in stool, the majority of Ethiopian studies have been solely performed by classical and biochemical identification methods.Both studies using molecular tools, included limited numbers of samples (less than 100), and each was carried out in a single study center [7,20].Also, studies involving humans, poultry, and/or cattle from the same households are rare, participate small sample sizes (total number of human and animal samples combined: n = 347 and 70) [20,21], and describe individuals of not age specific [20] and infants [21], regardless of the status of their diarrhea.
Campylobacter exposure to antimicrobials is expected to be high.Inappropriate empirical treatment of enteritis in children is common in the study area, and resistant infections result in toxicity, increased morbidity and mortality rates, and a heightened cost of care [12].Up till now, data on Campylobacter enteritis and antimicrobial susceptibility testing (AST) profiles in children with acute diarrhea in homes with backyard farm animals and/or companion pets is expected to be inaccurate in Ethiopia, due to the use of difficult to compare and/or suboptimal methods for isolation and AST.
Around 85% of rural households in Ethiopia own cattle, sheep, goats, poultry, and pets which interact closely with family members in the residences [22].They are frequently kept indoors and roam freely so that animal feces are present in-house.As a result, young children can acquire Campylobacter directly from animals or food, water and surfaces contaminated with animal waste.However, knowledge on the occurrence of Campylobacter between children under five and animals is lacking in Ethiopia.Due to the lack of toilet facilities, human defecation in the open, close to homes and drinking water sources, is also a common practice.Therefore, assessing the presence of Campylobacter in children and animals, along with their susceptibility towards locally available antimicrobials, is highly relevant.
Consequently, the objectives of the present study were to determine the occurrence of Campylobacter in a large cohort of children under the age of five years with diarrhea, as well as in backyard farm animals and companion pets associated with these children.Moreover, the study aimed to determine the AMR profiles of Campylobacter isolates from children and animals to antimicrobials locally prescribed or available and recommended by standard guideline [23].Based on these objectives, we conducted this study to test our hypothesis that due to environmental conditions, Campylobacter is a major cause of enteritis among children under five years old in Ethiopia, with contact with backyard farm animals are the primary source of transmission in the vicinity of Bahir Dar, northwest Ethiopia.New methodologies were used in order to generate baseline data that could support clinical decision-making and treatment, and provide insights on Campylobacter transmission and AMR between children and animals for the overall mitigation of campylobacteriosis and AMR.

Ethics statement
The Institutional Review Board of the College of Medicine and Health Sciences (CMHS) of Bahir Dar University reviewed and approved the study protocols with reference number CMHS/IRB 01-008.A formal letter of support was obtained from Bahir Dar University CMHS and Amhara Public Health Institute.Data and samples were collected after receiving written informed consent from the child's parents.Those formal consent Health centers were reached through the Bahir Dar City and Bahir Dar Zuria Health Departments.All survey, observation, and laboratory analysis data were anonymized, numbered, coded, and kept strictly confidential.For the proper management of patients, Campylobacter positive results were conveyed to nurses working at the health centers.Members of the households were informed about good practices in sanitation and hygiene to prevent Campylobacter infection and diarrhea.After authorization from the Ethiopian Food and Drug Authority, colonies presumptive of Campylobacter were transported from Bahir Dar, Ethiopia to Ghent, Belgium following national and international specimen transfer rules and regulations.

Study participants
Children under five suffering from diarrhea, who sought treatment at the health centers involved, and from whom sufficient feces for sample collection could be collected, were eligible to participate.Children were eligible to participate upon written parental consent.Children experiencing diarrhea for more than seven days and coming from a household without any backyard farm animals were excluded.received from the administrative offices, Kinbaba, Wonjeta, Yinesa and Robit Health Centers provided health care services for 32,668, 42,898, 17,303 and 28,401 people, respectively.Han, Zege and Abay Health Centers served 9,354, 5,025 and 10,306 people, respectively.From each of the health centers within a 0.4 to 12.8 KM radius, every household with the suspected cases was visited.

Stool sample and data collection from children
Samples were collected at each health center and 5-10 grams stool per child were transferred in a sterile plastic cup for storage [21], labeled with child code and date of collection.Samples were kept refrigerated at the health centers until transportation to the microbiology laboratory at Bahir Dar University within 4 hours of collection.Along with the sample, data on children's age, sex, residence, and clinical details (fever, abdominal pain, vomiting, duration of diarrhea, and frequency of passing stool per 24 h) were recorded from each child during face-to-face interviews with their parent(s).

Fecal sample collection from animals
Following the sampling of the children at the health centers, a visit to their respective homes was organized within one week.From the animals present at home, about 25 grams of fresh feces from bovines, sheep, and goats were collected using a clean spoon-attached stool cup [20].In case of lack of spontaneous defecation from the livestock, feces were collected from the rectum [25].For poultry, 10-15 grams of fresh cecal droppings were collected.Finally, the family's indoor garden was observed for freshly excreted feces from pets.A combination of pooled and subset sampling methods were employed, tailored to the circumstances of each animal species in various household scenarios.For poultry, bovine, sheep, and goats, pooled or subset sampling were employed, while calves and companion pets were sampled individually, as the latter were only few per household.

Isolation and preliminary identification of Campylobacter
Standard microbiological practices were followed for transportation, isolation, and preservation of isolates.From each stool and fecal sample, about 1 gram was homogenized in 10 ml of sterile physiological water and gently vortexed for 30 seconds.A polycarbonate filter of 0.6-μm pore size (Global Life Sciences Solutions Operations Ltd., Little Chalfont Buckinghamshire, UK) was placed onto a modified charcoal cefoperazone deoxycholate agar plate (Campylobacter Blood-Free Selective Agar Base, CM0739, Oxoid, Basingstoke, UK) [26].Six drops of the fecal suspension were added on top of a filter membrane and allowed to pass passively for 30 minutes at 37˚C, ambient air [27].After filtration, the filter was removed, and plates were incubated in a glass jar for 48 to 120 h at 37˚C in a microaerobic atmosphere (burning candle) [16,28].Plates were checked after 48, 72, 96, and 120 h of incubation for the growth of typical Campylobacter colonies.
Preliminary identification of Campylobacter isolates was made by typical colony appearances on mCCDA plates.Translucent, grey-white with a metallic sheen, watery and spreadlike droplets to a continuous film (swarming) colony, were considered presumptive Campylobacter colonies [29].
Subsequently, after microscopic confirmation of the presence of fine and spiral rods by safranin staining [14,26], typical colonies were streaked onto a new mCCDA plate, and incubated for 24-72 h at 37˚C, under microaerobic conditions.Performance of mCCDA, at 37˚C, and microaerobic incubation conditions were tested using Campylobacter reference strains from the Belgian Coordinated Collections of Microorganisms (BCCM/LMG).

Storage and transportation of presumptive Campylobacter isolates
Upon growth, pure colonies with typical Campylobacter morphology were swabbed from the second mCCDA plates with a sterile cotton swab, transferred to cryotubes (Thermo Scientific, Suxhou, Jiangsu, China) with 1 ml sterile sheep blood and stored at-80˚C.An additional aliquot was kept at -20˚C in Eppendorf tubes with 1 ml of sterile physiological water.The presumptive Campylobacter isolates were transported on dry ice from Ethiopia to the Microbiology Laboratory (LMG) at Ghent University, Belgium.

Re-cultivation of presumptive Campylobacter isolates
In Belgium, the suspensions were sub-cultured on blood agar (BA) plates (Columbia Agar Base, CM0331, Oxoid, with 5% defibrinated horse blood (E & O Laboratories Limited, Burnhouse, Scotland)) [28,30].In addition, 0.5 ml of each cell suspension was transferred to 10 ml of Preston Broth (PB) (Nutrient Broth No. 2, CM0067, Oxoid, with 2ml of sterile distilled water and Modified Preston Campylobacter Selective Supplement, SR0204E, Oxoid, and 5% lysed horse blood) in a sterile screw-capped bottle [31].Blood agar plates and PB tubes were incubated at 37˚C in microaerobic conditions through the introduction of a gas generator to provide a mixture of gases (10% CO2, 10% H2, and 80% N2) into metal jars.After 24 h of incubation, 10 μl of each enrichment broth was spread onto a BA plate and incubated for 48 h as described above [26].The remaining PB tubes were incubated again for 24h.Of samples for which no growth was obtained after 24 h, another 10 μl was spread from the corresponding PB culture onto a BA plate and incubated for 48-120 hrs [31].Typical Campylobacter growth on the BA plates was harvested for further confirmation and identification at the species level using MALDI-TOF MS.

Campylobacter species identification by MALDI-TOF MS
Typical Campylobacter colonies from BA plates were analyzed, mass spectra (MSP) generated and identified by MALDI-TOF MS (Bruker's MALDI-TOF Microflex LT/SH LMGE-0422) as described previously [32,33].Interpretations of results were made by comparing the generated MSP Bruker log scores to the Bruker database MSP-10833 and LM-UGent iD MSP-4778 database using the Bruker Compass software (Bruker Daltonics).To ensure that the run was valid, calibration was performed before each set of measurements using 1μl Bacterial Test Standard (BTS) (Bruker Daltonics).Identification of Campylobacter at the species level was considered if scores were � 2.2, according to the Bruker MALDI Biotyper specifications.Isolates that could not be re-cultivated, or yielded a score < 2.2 identified, were further analyzed by PCR.

Antimicrobial susceptibility testing
The antimicrobial susceptibility of C. jejuni and C. coli to ciprofloxacin, doxycycline, erythromycin, and azithromycin was determined using the Epsilometer gradient strip diffusion method (ETEST) (bioMe ´rieux, Marcy-l'Etoile, France) on Mueller-Hinton Fastidious (MHF) agar (bioMe ´rieux, Marcy-l'Etoile, France).Inocula were prepared by suspending colonies in MHF broth to a 0.5 McFarland density using a nephelometer.Inoculated plates were incubated under micro-aerophilic conditions using a gas mixture of 6% O2, 7.1% CO2, 7.1% H2, and 79.7% N2 (Anoxomat) at 37˚C.After 48 h of incubation, the MIC values were read and interpreted based on the 2023 EUCAST Clinical Breakpoint Tables V. 13.0 [23].Campylobacter jejuni ATCC 33560 quality control strain was included for each separate run.Antimicrobials were selected based on availability and prescription in the study area, and standard treatment and AST guidelines.Ciprofloxacin is the first choice of treatment for enteritis in the study area.In addition, ciprofloxacin and azithromycin are listed as first and second choices for enteritis treatment in primary care.All four antimicrobials tested are locally available.

Statistical analysis
GPS coordinates of the visited homes were recorded, and their spatial distributions were visualized using ArcGIS version 10.5 software.Data were analyzed using IBM SPSS Statistics for Windows, Version 29.0.Armonk, NY: IBM Corp. SYSTAT.Baseline characteristics of the children were summarized using numbers with percentages for categorical variables.Pearson Chisquare tests were used to observe a significant association between categorical variables (gender, different age groups, frequency of stool per day, and duration of diarrhea) and Campylobacter infection in children.A Games-Howell post-hoc test was used to identify among which specific age groups the significant difference in Campylobacter infection occurs.
The presence of significant association between each symptoms related to diarrhea (watery diarrhea, foul-smelling diarrhea, vomiting, abdominal pain, and fever) and Campylobacter infection was evaluated using the chi-square test.Fisher's exact test was used to observe the association between Campylobacter infection and bloody stools.
Both the Shapiro-Wilk and Kolmogorov-Smirnov tests were used to test the normality of continuous data such as age of the children in months, duration of diarrhea in days, frequency of diarrhea per day, and number of siblings per household.All these data were not normally distributed and are reported with medians and interquartile ranges (Q1 to Q3).The median ages of children with and without Campylobacter infection were compared using the Mann-Whitney U test.Likewise, the Mann-Whitney U test was used to compare the median duration of diarrhea, the median frequency of diarrhea, as well as the median number of siblings, between children with and without Campylobacter infection.
Finally, a chi-square test was computed to assess differences in the rate of resistance to ciprofloxacin between C. jejuni and C. coli isolates.Likewise, the rate of resistance to doxycycline between C. jejuni and C. coli was compared using the chi-square test.P-values less than 0.05 were considered statistically significant.

Occurrence and identification of Campylobacter
As shown in Table 1, 303 children from 303 different households participated in the study and provided a fecal sample.In addition, fecal samples from 253 poultry, 213 bovines, 105 calves (under 1 year of age), 99 sheep, 22 dogs, 17 goats and 2 cats were collected.Upon first screening in Ethiopia (preliminary identification), Campylobacter was presumptively present in 250 of the 1014 samples.After shipping to the lab of microbiology, Ghent University, Belgium, presumptive isolates from 34 samples could not be cultivated or identified by PCR.The remaining isolates from 129 samples could be recultivated and resulted in 135 identifications as Campylobacter species by MALDI-TOF MS, and an additional 97 identifications via PCR in 87 samples for which re-cultivation was not possible.From these confirmed 216 positive samples, 200 had one type of Campylobacter species, and 16 had co-infection with both C. jejuni and C. coli resulting in the identification of 232 Campylobacter isolates.The majority of the isolates were C. jejuni (n = 194) followed by C. coli (n = 34).Additionally, four animal isolates were identified as C. fetus (n = 2) and unspecified Campylobacter spp.(n = 2).
In total, 59 of the isolates originated from the stools of 59 children (out of 303 samples), and 173 isolates from 157 animal fecal samples (out of 711 samples).The occurrence of Campylobacter in animals varied per host, with rates up to 64% in dogs to 6% in bovines (Table 1).In animals, co-occurrence of C. jejuni and C. coli was observed in dogs (9%), poultry (4%), sheep (3%), and calves (1%) (Table 1).

Demographic and clinical details of children with diarrhea and Campylobacter infection
Out of the 303 under five children who participated in this study, the exact age of two children missed being recorded.The median age of the 301 children was 19 months with an interquartile range of 12-31.5, and children with Campylobacter infection were significantly younger (13 months) compared to those without (19 months) (Mann-Whitney U test: p < 0.001).

Occurrence of Campylobacter at the household level
Out of 303 households, 245 households could be visited for the subsequent collection of 711 additional animal fecal samples.The median (Q1, Q3) number of animals included per household was 3 (2,4).Campylobacter was present in at least one host in 149 (61%) of 245 households.In 27 (18%) of these, Campylobacter was only detected in the child, while in 23 (15%) it was detected in both a child and an animal species (Fig 4A Campylobacter was detected in animal species but not in children in 99 households.Among these, 89 were detected in a single animal species, with 68 (76%) from poultry, 7 (8%) from calves, 5 (6%) from dogs, 5 (6%) from sheep, 3 (3%) from bovines, and 1 (1%) from goats
The proportion of C. jejuni isolates resistant to ciprofloxacin was 3 (9%) in children and 10 (19%) in poultry.The proportion of C. coli isolates resistant to ciprofloxacin was 3 (75%) in ruminants and 9 (47%) in poultry.All C. coli isolates (n = 5) resistant to doxycycline were from poultry.Of the 128 isolates tested, only 1 (1%) from a child was resistant to azithromycin and 1 (1%) from poultry to erythromycin (Table 3).
The MICs of ciprofloxacin ranged from 0.008 to > 32 mg/L in C. coli and from 0.002 to > 32 mg/L in C. jejuni.The MICs of doxycycline ranged from 0.03 to 64 mg/L in C. coli, and from 0.016 to 256 mg/L in C. jejuni.Ciprofloxacin has a bimodal distribution of MICs for C. coli isolates (0.008-0.125 mg/L versus 4-32 mg/L).The MICs for doxycycline are also bimodal (Table 4).
As shown in Table 5, of the C. jejuni isolates simultaneously isolated from a child and poultry in 4 households, all were concurrently susceptible to azithromycin, doxycycline and erythromycin.However, in one household, a poultry isolate was resistant to doxycycline, and in another household, a poultry isolate was resistant to erythromycin.For ciprofloxacin, two isolates were concurrently susceptible to increased exposure, and two were resistant in poultry but not in children.In two households where C. jejuni was isolated from dogs and poultry, isolates from both sources were susceptible to erythromycin and azithromycin.However, in one household, an isolate from a poultry was resistant to ciprofloxacin, and in another household, a poultry isolate was resistant to doxycycline (Table 5).Furthermore, from three households, where C. jejuni and C. coli were isolated from a single poultry sample, both species were susceptible to all antibiotics.Similarly, from a single calf in a household, both C. jejuni and C. coli isolated were susceptible to all antibiotics (Table 5).

Discussion
This study presents the occurrence, spatial distribution, and AMR of Campylobacter in a large cohort of under five children with acute diarrhea and various animal species.Using a membrane filtration culture technique and candle extinction microaerobic incubation system, we successfully isolated thermotolerant Campylobacter.MALDI-TOF MS and PCR confirmed 86% of presumptive isolates as Campylobacter.Compared to earlier studies in Ethiopia, the rate of infection in children is high, with children aged 6 to 11 months having the highest infection rates.Co-occurrences of Campylobacter in both children and animals sharing the same home may suggest possible household-level transmission.The spatial distribution suggests that the vicinity of Bahir Dar may be a hotspot area for the presence of Campylobacter.Antimicrobial resistance profiles showed low resistance to erythromycin and azithromycin but higher rates for ciprofloxacin and doxycycline, particularly in poultry.The findings could be applicable to regions with similar socio-economic and cultural contexts, where backyard farming is common.The fastidious nature of Campylobacter challenges the reliability of stool culture for estimation of the burden of Campylobacter in regions with restricted resources for laboratories.Previous studies in Ethiopia lacked consistency in culture techniques impeding a comprehensive understanding of Campylobacter epidemiology.To address this gap, the current study, the first of its kind in Ethiopia, adopted a membrane filtration culture technique that resulted in the successful isolation of thermotolerant Campylobacter from children, backyard farm animals and dogs.A significant portion of presumptive samples were confirmed as Campylobacter via molecular tools.Out of the 232 Campylobacter identifications, 58% were identified using MALDI-TOF MS, and 42% via PCR.The cells which could not be resuscitated and recovered by culture but that were identified via PCR, could be the viable but non-culturable (VBNC) cells [37], or injured cells due to freeze-thaw during an unstable power supply, or the lack of representation in the MALDI-TOF MS database.Moreover, presumptive isolates in a subset of samples not identified as Campylobacter by molecular tools could be bacteria such as Proteus and Acinetobacter, which resemble Campylobacter morphology on mCCDA media [38], or emerging Campylobacter species.
Compared to earlier studies in different regions of Ethiopia, the rate of Campylobacter infection in children (20%) in the present study, exceeded the rates reported in southwest (9-17%) [39,40], southern (7-13%) [41,42], and northwest (6-15%) [43][44][45] Ethiopia.Previous studies in Ethiopia typically used selective Campylobacter agar without filtration, and incubation at 42˚C for 24 to 48 hours.A study in eastern Ethiopia reported a detection rate of 15% in infants by PCR [7].However, another study from southern Ethiopia [21], reported a detection rate of 49% in infants using selective chromogenic agars.The rate of infection in children is also higher than previously reported rates in Africa [8], Asian countries [46,47], Ecuador [48] and Poland [49].The variation in the findings could be due to the stool culture techniques used to isolate and identify Campylobacter.The studies in Africa [8,39,40,42,44,45] applied biochemical tests that have identification limitations stemming from the low metabolic activity of Campylobacter [30].Thus, the membrane filtration culture method, combined with specific incubation conditions is an effective and easy method for screening C. jejuni/C.coli in routine diagnosis, and research in settings lacking organized laboratories.
Differences in recruitment methods for including diarrheic children could also contribute to variations in Campylobacter occurrence.A study in northwest Ethiopia [43] recruited diarrheic children through house-to-house searches, whereas the present study enrolled children seeking medical care for diarrheal illness.This is because many symptomatic instances of Campylobacter might not seek medical attention.Differences in frequency and duration of diarrhea, admission status, and considerations for immunodeficiency and malnutrition across studies, could contribute to variations in Campylobacter occurrence among children.In the present study, children with a minimum of three episodes of diarrhea per day lasting for 1 to 7 days were included, while studies in northwest Ethiopia [43,45] enrolled those with watery diarrhea once daily for 1 to more than 14 days.Studies from Asia enrolled admitted children [46,47], whereas outpatient children were included in the present study.Additionally, a study in Lebanon [46] excluded children with immunodeficiency, malnutrition, and multiple malformations, a factor not accounted for the present study.Furthermore, children from urban and rural settings may have different transmission intensities of Campylobacter.Factors such as differences in sanitation and hygiene levels and nutrition between studies could contribute to observed variations in the occurrence of Campylobacter in children.
In the present study, the number of Campylobacter infections was significantly higher in children aged 6 to 11 months compared to those aged 36 to 47 months and 48 to 60 months in line with previous studies in Ethiopia [45], and elsewhere in the world [8,50,51].As reported in Ethiopia [39,40,45], higher infection rates were observed in children below three years old in the present study (p = 0.012).This can be attributed to differences in the development and maturation of intestinal microbiota among the different age groups [52].The intestinal microbiota of the children is more mature from the age of two years onwards [53,54].Earlier exposure to Campylobacter also induces specific partially protective immunity.Besides, children from 6 to 11 months are more likely to be exposed to contaminated surfaces, foods, and water after weaning, and have less developed immunity [8,52].
In our study, C. jejuni prevailed C. coli in both children and animal species consistent with what is reported before [19].This could be due to the diverse nature of C. jejuni, enabling its widespread colonization in mammals and poultry [55].Moreover, C. jejuni has better virulence factors and adaptation mechanisms that enable it to survive and proliferate in the human gastrointestinal tract compared to C. coli [55].The observed occurrence of C. jejuni in children (19%) in the present study exceeded previously documented rates in Ethiopia (9 to 12%) [40,56] as well as rates reported among children in the African context (5.2-12.3%)[8].The rate of C. jejuni in ruminants (5-8%) is comparable with previous reports from other parts of Ethiopia (ranging from 3% to 9%) [7,18,57].A higher rate of C. jejuni has already been reported from the feces of poultry (81%) and cattle (54%) in southwest Ethiopia [58].Methodological differences could contribute to the observed variations, as several earlier studies in Ethiopia relied on biochemical tests [18,41,57,58].
The overall rate of C. coli (3%) in the present study fell within the spectrum of rates (ranging from 1% to 11%) reported in previous studies mentioned above [18,20,40,56,58].The isolation rate of Campylobacter coli was higher in animal species than in children in the present study.The strains of C. coli prevalent in animals might be less adapted to infect and cause disease in children.Additionally, the simultaneous excretion of both C. jejuni and C. coli was observed in backyard farm animals and companion pets but not in children.This could be due to competition with the intestinal microbiota, and variations in exposure to pathogens, body temperature, strain virulence, and immune responses [26,59].
Apart from a study in central Ethiopia reporting an 8% prevalence of C. fetus from livestock [20], other studies in Ethiopia did not isolate C. fetus from fecal animal samples.The use of an incubation temperature of 42˚C inhibits the growth of most non-C.jejuni and non-C.coli Campylobacter spp.The presence of C. fetus only in one bovine and one calf stool sample and the inability to isolate other slow-growing Campylobacter species in the present study might be linked to the absence of hydrogen gas in the candle extinction microaerobic conditions during isolation [27,60].
Regarding animal species, the occurrence of Campylobacter spp. was higher in dogs and poultry compared to sheep and cattle in concurrence with a finding from southwest Ethiopia where the percentage of isolation was higher in chicken compared to sheep and cattle [58].Conversely, other studies in the central and southwest regions of Ethiopia reported higher rates of C. jejuni in ruminants and C. coli in sheep [20,58,61].The isolation rate of Campylobacter among household dogs (64%) was higher compared to rates from other countries (11-45.4%)[62,63], possibly due to the unrestricted movements of dogs in the neighborhood, potentially increasing their exposure to various sources of contamination, including the excreta of children and from consumption of dairy products [64].Similarly, the isolation rate of Campylobacter in poultry (44%) was higher than a report from central Ethiopia (13%) [20].Relatively lower rates of Campylobacter isolation in sheep (9%) and cattle (6-9%) were documented compared to previous reports in Ethiopia (11%-38% in sheep and 12.6-12.7% in cattle) [18,57,58].However, considering the fact that 85% of households in Ethiopia own backyard farm animals [22], and keep them in family houses, and the concurrent excretion of C. jejuni and C. coli in a few cattle in the current study highlights the need for further research.Differences in animal species, age, and handling practices and geographic regions may be attributed to the observed differences in Campylobacter occurrences among backyard farm animals.
Investigating Campylobacter occurrence between children and animals sharing the same house is crucial for designing targeted interventions.Campylobacter was detected in 61% of the family's homes visited, which is higher than a previous study from central Ethiopia (42%) [20].This could be due to variations in the age and diarrhea status of human participants.We screened stool samples collected only from children under five with diarrhea seeking medical care at health centers, while the study in central Ethiopia collected pooled human fecal samples at the household level regardless of both diarrheal status and age specifications.
The occurrence of joint cases of C. jejuni among children and animals (for instance, childpoultry in 14 houses, child-bovines in 2, and child-dog-poultry in 2 homes), along with the matched AST profiles of C. jejuni from children and animals in two homes, may suggest the presence of possible household-level transmission of Campylobacter via exposure to floors and various surfaces contaminated with feces.This is strengthened by the free access of animals to the sitting, kitchen, and dining areas of the families, as observed during data collection.This finding is in accordance with other studies [20,48,65].However, the detection of Campylobacter exclusively in children in some homes or exclusively in animals in other homes, as well as the unmatched AST profiles of C. jejuni isolates from children and animals in three homes, could indicate the presence of different contamination sources [65].These indicate the complexity of the transmission dynamics of Campylobacter.
Knowledge on the AMR of Campylobacter isolates from children, backyard farm animals, and companion pets is scarce in Ethiopia.The present study determined, for the first time in Ethiopia, MICs for the antimicrobials recommended to treat human campylobacteriosis.Only 0.8% of tested C. jejuni isolates were resistant to erythromycin and azithromycin, which is consistent with recent CDC, European Union (EU), and global reports [49,66,67].However, a MIC of azithromycin (� 256 mg/l) was observed in a C. jejuni from a child.This could be due to the unrestricted use of azithromycin for mild pneumonia in the study area.Higher rates of resistance to erythromycin were reported in Campylobacter isolates from young children in southwest Ethiopia (11-18%) [39,40], and northwest Ethiopia (28%) [44] and in poultry in central Ethiopia (15%) [20].These studies employed disc diffusion techniques for susceptibility testing.
The 20% rate of ciprofloxacin resistance is consistent with findings in Europe (18-22%) [59,68].However, variable resistance rates to ciprofloxacin were reported in isolates from children in the southwest (5.3% to 15.8%) [39,40], southern (55%) [41] and northwest (16%) [44] Ethiopia.An 8% rate of resistance in bovines and 13% combined (human and backyard farm animals) were also reported in Ethiopia [18,20].C. coli displayed significantly higher resistance to ciprofloxacin than C. jejuni (p < 0.001), consistent with reports from EU [67].The high rates of ciprofloxacin resistance among C. jejuni isolates from poultry and C. coli isolates from calves (67%) in the present study is concurrent with a study from Poland for C. jejuni [69] and a report from EU countries for C. coli [67].As only a few isolates from ruminants were tested in this study, caution needs to be taken in the interpretation of this finding.The MIC of ciprofloxacin showed a bimodal distribution in C. coli compared to a monomodal one in C. jejuni indicating different resistance mechanisms or the presence of subpopulations in C. coli.The MICs for doxycycline are also bimodal and in agreement with the EUCAST epidemiological cut-offs (ECOFFs).
The overall rate of Campylobacter resistance to doxycycline (11%) was lower compared to worldwide reports (28-42%) [70,71].The rate of resistance was considerably higher in C. coli than in C. jejuni, consistent with reports from the EU [67].The rate of doxycycline resistance in C. jejuni was higher among isolates from poultry than children consistent with the study from Poland [69].The 8% ciprofloxacin-doxycycline co-resistance observed in Campylobacter isolates in the present study is in line with the 6% and 9% rates reported in U.S., [72] but lower than the 15% co-resistance rates in Bulgaria [68].Almost 90% of ciprofloxacin-doxycycline coresistance was observed among isolates from poultry.The overall co-resistance was significantly higher in C. coli isolates than C. jejuni, consistent with the report from EU [67].
This study presents the matching between susceptibility profiles of isolates from children and animals (Table 5).The presence of matched AST profiles in two homes was confirmed, while unmatched in three homes.In general, the resistance of Campylobacter to ciprofloxacin and doxycycline in the present study may be due to the common use of these antimicrobials for treating enteric infections in the study area.The predominance of resistance in isolates from backyard poultry where Campylobacter can easily multiply, survive in droppings for up to 6 days after excretion [73] and spread to children, humans, and companion pets is of particular concern.The observed resistance in poultry may be due to specific poultry-associated mechanisms, reverse zoonotic transmission, acquired resistance from wild birds, local farming, or environmental sources, rather than selection pressure, as antimicrobials are not used in backyard poultry in this particular setting.Moreover, the unmatched AST profiles of C. jejuni isolates across poultry and children, and poultry and dogs observed in a few homes (Table 5) support these explanations.
Assessment of the occurrence of Campylobacter using large sample sizes at the children-animal interface, the use of membrane filtration culture technique for the first time in Ethiopia, as well as successful transportation from Ethiopia to Belgium on dry ice, and subsequent recultivation under microaerobic conditions by introducing a gas mixture, were the strengths of this study.In addition, the confirmation of isolates using MALDI-TOF MS and determination of AMR based on gradient strip diffusion susceptibility testing is the first for Campylobacter from Ethiopia.
There are also several important limitations that should be considered when interpreting findings from this analysis.The present study does not use enrichment before filtration-based plating, which may have limited the isolation of Campylobacter that were present in low numbers or under stress in the tested matrix.While the method employed favorably isolated C. jejuni and C. coli, which are responsible for the majority of gastrointestinal diseases related to Campylobacter infection, it may have overlooked slow-growing emerging Campylobacter species such as C. lari, C. hominis, C. upsaliensis, C. hyointestinalis, C. curvus, C. concisus, and others due to the lack of a hydrogen-enriched atmosphere during the initial isolation microaerobic incubation.Additionally, fast-growing Campylobacter may also be missed in some samples.Furthermore, the use of candle extinction microaerobic incubation may have a lower isolation rate compared to methods involving a gas generator to provide a mixture of gases (10% CO 2 , 10%H 2 , and 80% N 2 ).
Moreover, the cross-sectional nature of the study and the lack of genetic analysis limit to elucidate the transmission dynamics of Campylobacter between animal species and children, even in households where C. jejuni co-occurs among both.Additionally, due to logistical constraints and feasibility, environmental samples such as surfaces, dust from the living areas of children and animals, and water sources were not examined.This narrows the discussion on the role of the environment in transmission or as an opportunity for intervention.As a result, the present study primarily addresses the human and veterinary domains of the One Health doctrine.

Conclusions
Campylobacter jejuni is prevalent in children below the age of 3 years with diarrhea in Ethiopia.Its abundance in children, shedding from poultry and companion pets and concordant presence in both a child and animal species with matching AST profiles in a few homes may suggest possible household-level transmission between animals and children.Based on the findings, treatment of severe campylobacteriosis in children should be guided by E-test susceptibility testing.In order to preserve efficacy and minimize the risk of AMR development, azithromycin and erythromycin should only be used for the treatment of severe campylobacteriosis that is resistant to ciprofloxacin.The observed high prevalence of Campylobacter in children and co-occurrences among children and animals and animal species within households suggest further genetic analysis and longitudinal studies to comprehensively understand Campylobacter transmission, mechanism of AMR, and ascertain causality to children.This finding highlights the need for targeted interventions such as raising awareness of transmission routes, improving hygiene, and proper handling of animals and disposal of their waste.Clinicians, veterinarians, and policymakers should integrate these findings for evidence-based diagnosis and treatment, further research, surveillance, and mitigation of Campylobacter transmission.

A
Fig 1. (A) Location map of the study area.(B) spatial distribution of health centers and sampled households.N.B: Shapefiles for (A) and (B) are accessed from https://www.naturalearthdata.com/.GPS coordinates of the health centers and households plotted in (B) are collected from each health center and home during sample collection.https://doi.org/10.1371/journal.pntd.0012241.g001 Likewise, C. coli was isolated most from dogs (14%), then from poultry (8%), sheep (4%), calves (3%), children (1%), and bovines (1%).The occurrence of Campylobacter is shown in Fig 2. Of the 1014 samples examined, C. jejuni was present in 194 (19%) and C. coli in 34 (3%), with 16 out of 711 samples from animal origin only (2%), with co-occurrence of C. jejuni and C. coli.Campylobacter fetus was isolated from one bovine and one calf fecal sample.In total at least one Campylobacter species was present in 216 of the 1014 samples examined (21%).

Fig 3 .
Fig 3. Campylobacter infection per age group of children.Based on Games-Howell post hoc test multiple comparisons, the proportion of Campylobacter infection was significantly higher in age groups from 6 to 11 months compared to 12 to 23 months (p = 0.05), 36 to 47 months (p = 0.006) and 48 to 60 months (p = 0.004).No significant difference was observed between the agegroup 6 to 11 months and the remaining age-groups (p > 0.05).https://doi.org/10.1371/journal.pntd.0012241.g003

As visualized in Fig 5 ,
the majority of households sampled were from Kinbaba and Han health Center catchment areas.Households positive for Campylobacter are observed across the radius of each health center, with a higher concentration in Kinbaba, followed by Han.For children, Campylobacter positivity was concentrated in the Kinbaba and Han catchments, while positivity in poultry seemed evenly distributed across the radius of each health center (Fig5A).At species level, C. jejuni exhibited an even distribution, while C. coli appeared scattered.A single instance of C. fetus was detected in Han and Yinesa (Fig5B).

Fig 5 .
Fig 5. Spatial distribution of households with and without Campylobacter.Spatial patterns of Campylobacter per host in (A).Spatial patterns of Campylobacter species per sampled households in (B).N.B: Shapefiles for both (A) and (B) are accessed from https://www.naturalearthdata.com/.GPS coordinates of the health centers and the households plotted are collected from each health center and home during sample collection.In panel (A), households delineated in green indicate those that tested negative for Campylobacter, while other colours represent households that tested positive per host.In panel B, grey delineated households indicate those that tested negative for Campylobacter, whereas different colours represent positive households per Campylobacter species type.https://doi.org/10.1371/journal.pntd.0012241.g005

Table 1 . Campylobacter occurrence and species distribution in children, backyard farm animals and companion pets. Number of hosts Campylobacter occurrence Campylobacter species distribution
N: number https://doi.org/10.1371/journal.pntd.0012241.t001PLOSNEGLECTED TROPICAL DISEASES

Table 2 . Demographic and clinical details of children under five with diarrhea with and without Campylobacter.
). Campylobacter was detected in a child and poultry in 14 of these 23 households, as well as in child-bovine, child-dog-poultry, child-sheep, child-bovine-dog, and child-calf-poultry combinations in 2, 2, 1, 1, and 1 households, respectively (Fig4B).The majority (19/23) of these co-occurring identifications were consistent at Campylobacter species level between the different hosts.More specifically, 18 were consistent between hosts and were identified as C. jejuni, and one identified as C. coli.In four cases, different species of Campylobacter were identified in different hosts (children and animals) from the same household (Fig4B).
, MIC values were read and interpreted for 128 Campylobacter isolates (103 C. jejuni, 24 C. coli, and one C. fetus).Isolates demonstrated resistance rates of 20% and