Prevalence and Risk Factors Associated with Campylobacter Infection in Diarrheal Patients in Busia County, Kenya

Introduction: Campylobacter are zoonotic bacteria that cause gastroenteritis worldwide with the species, Campylobacter jejuni and Campylobacter coli commonly associated with human diarrhea. Transmission is mainly through direct contact with farm animals, consumption of chicken and contaminated water. There is paucity of data on the epidemiology of Campylobacter in developing countries despite its global widespread and expansion of poultry farming; hence there is the need to explore and build on the available data. This study aimed at determining prevalence and homestead risk practices associated with Campylobacter infection in diarrheal patients in Busia County. Methods: A cross-sectional study was conducted from February, 2017 to April, 2019. Stool samples were collected from patients of all ages attending Busia County referral Hospital and structured questionnaires on homestead associated risk practices administered. Isolation and identification of Campylobacter species was performed using standard culture method on Modified Charcoal Cefoperazone Deoxycholate medium and confirmed by mPCR. Factors associated with Campylobacter infection were evaluated using logistic regression analysis. Results: A total of 132 (11.6%) Campylobacter comprising 89.2% C. jejuni and 10.8% C. consumption of raw milk (OR 7.14, 95% CI: 1.96 - 18.24, P = 0.001) posed risk among those aged > 5 years respectively. Conclusion: Campylobacter jejuni is the leading cause of Campylobacter infections in diarrheal patients. Personal hygiene awareness of mothers/caregivers and proper animal husbandry especially where livestock-human interaction is common are important practices which require the County government sup-port. Further studies are required on sex specific age difference, other social economic factors, domestic animals and the role played by the environment in the transmission of Campylobacter infection. These would advance knowledge and understanding on source attribution and transmission dynamics for effective control and management of the infection.


Introduction
The genus Campylobacter is a zoonotic foodborne pathogen that causes diarrheal disease; the two major species are Campylobacter jejuni and Campylobacter coli [1]. Campylobacter jejuni is responsible for over 90% of human campylobacteriosis and is frequently isolated from poultry (81%) while C. coli causes less than 10% of human disease and is most frequently isolated from pigs (100%) [2] [3].
Campylobacteriosis is a problem and an economic burden to human population worldwide with an estimated 96 million cases annually, exceeding those of salmonellosis, shigellosis and E. coli: 0157H7 combined [4]. It is the third most common cause of hospitalization globally after Rotavirus and Salmonella infections and is associated with the loss of up to 7.5 million disability-adjusted life  [6]. Human Campylobacter infections are categorized into gastrointestinal conditions, including inflammatory bowel diseases (IBD), diarrhea, Barrett's esophagus and colorectal cancer [7] and extragastrointestinal manifestations including bacteremia, lung infections, brain abscesses, meningitis, and reactive arthritis [8]. Incubation period is 1 -10 days and clinical symptoms range from acute abdominal pain, watery diarrhoea, fever, nausea, vomiting, headache and muscle pain [9] which may persist for a week or longer [10]. The severity of illness varies from mild malaise to dehydration sufficient to warrant hospitalization and can be life-threatening [11].
Although campylobacteriosis is largely perceived as a foodborne infection with consumption and handling of poultry and poultry-related products being the major source of all Campylobacter infections [12], there is increasing evidence of additional routs of transmission including direct contact with other farm animals, consumption of undercooked foods and contaminated environment [13] [14]. Wild birds also play a role as reservoirs and can spread Campylobacter to humans through the environment [15]. Human-to-human transmission has been observed, although at low frequencies [16].
The epidemiology of Campylobacter enteritis differs between high-and low-income countries [17]. In high-income countries, clinical signs are symptomatic and occur in all age groups [18] whereas in low resource settings, most cases are asymptomatic with young children being more predisposed due to underdeveloped immunity. Adults acquire a level of protective immunity following repeated exposure and become predisposed when immunity gets low at senior age. Globally, the distribution of Campylobacter is attributed to asymptomatic colonization of the intestinal tract in a wide range of livestock species, use of animals manure, drinking contaminated water and poor hygiene conditions [19] [20] [21]. Studies have suggested that the incidence of Campylobacter enteritis is on the rise globally [22]. This increase has been noted much more in developed countries due to improved awareness and advances in detection practices [23].
In low and middle income countries (LMICs) such as Kenya, sources of Campylobacter infections have been less well documented due to lack of awareness and advances in detection practices to undertake national surveillance programmes and build a National epidemiological data banks of the disease. However, the available data indicate that the infection is endemic and on the rise [23]. The limited reports on Campylobacter are estimates from laboratory-based surveillance studies of pathogens responsible for diarrhoea [24] [25]. In Tanzania, the prevalence of human Campylobacter enteritis was reported at 21.6% [26] and between 12% -12.6% in Kenya [27]. Because of the importance of this disease and the rise in poultry farming in Kenya, this study was carried out with an aim of determining the prevalence and homestead risk practices associated with Campylobacter infection in diarrheal patients in Busia County, a region marked by high prevalence of HIV and proliferation of small-holder livestock production with over 95% of households keeping poultry [28].  [29]. The region has the highest rural human and livestock population densities and interaction of human and livestock in East Africa, operating in a mixed smallholder livestock production system [28]. Most households keep livestock, with the most common species being indigenous chickens. In addition to high rates of household livestock ownership, the community commonly slaughter and prepare chicken meals at home and rarely buy chicken from the few available retail outlets consequently leading to high exposures and transmission of zoonotic diseases [28] with regards to urban population, only Busia and Malaba towns in the County meet the minimum population threshold of 10,000 people for an urban center as per the section 8 and 9 of the Urban Areas and Cities Act 2011.

Study Population
The study population consisted of all patients presenting with diarrhea who sought treatment at the outpatient department at Busia County Hospital from February 2017 to April 2019.

Recruitment Strategy
After obtaining consent/assent from the adult participants and guardian/child, a systematic recruitment of the diarrhoeal (defined as passing of loose stool at least 3 times a day) patients [30] was conducted by requesting for stool sample in a clean plastic poly pot or rectal swab from children who could not pass stool. Semi-structured questionnaire was administered to obtain data on exposures in the 10 days before the onset of symptoms. The questionnaire comprised several sections, each representing different exposure groups and was pre-tested before the main recruitment for consistency. The interview was conducted in Kiswahili, English or Luhya depending on the language that was best understood by the respondent. Parent/guardians of the participating children were asked to complete the questionnaire on behalf of children.
The exposure variables were chosen on the basis of biological plausibility and previous knowledge on the risks associated with Campylobacter infection in other settings [13] [20] [21], and included socio-demographic factors such as age, residence, educational level and occupation. The occupation was divided into business which referred to those who earned livelihood by engagement in income generating activities such as working in hotels, local markets, shops and public transportation; the employed which referred to those who were salaried and worked in government, companies or NGOs; the unemployed which referred to those without any work with consistent income and finally those involved in poultry farming as their main upkeep; Behavioral and socio-economic factors included regular usage of toilet facilities within the homestead, types of foods commonly eaten and where purchased, exposure to infected diarrheal person(s) within the same household, pets and local and or international travel, types of houses lived in, source of cooking and lighting energy. The clinical section comprised of symptoms of Campylobacter enteritis and status on expanded immunization program (EIP) vaccine, any long term medication and immunosuppressive treatment based on a 4-week history.

Inclusion Criteria
Diarrheal patients of all ages seeking treatment at the outpatient department, Busia County Hospital who consented/assented to participate in the study and who self-reported to have not been on antibiotics for the previous 5 days.

Exclusion Criteria
Non-diarrhoeal cases and those that declined to participate.

Specimen Collection and Isolation of Campylobacter Species
Fresh stool/rectal swab samples were collected into a clean plastic poly-pot from diarrheal patients and transported in Amies media within 2 hours in cool box to the Centre for Infectious and Parasitic Diseases Control Research (CIPDCR), KEMRI, Busia laboratory for Campylobacter isolation. Standard microbiological techniques were used to culture and isolate Campylobacter jejuni and Campylobacter coli as described before [31]. Briefly; a gram of faecal sample or rectal swab was cultured on Exeter enrichment broth containing nutrient broth supplemented with lysed horse blood and incubated at 42˚C aerobically for 24 hours. A sub-culture was done by passing 100 µl of the enrichment broth culture on 0.45 µm cellulose nitrate filter onto modified Cefoperazone Charcoal Deoxycholate Agar (mCCDA) supplemented with SR155 (Oxoid) and incubated at 42˚C ± 1˚C for 48 ± 4 h under microaerobic conditions (5% O 2 , 10% CO 2 , and 85% N 2 ) generated by CampyGen TM (Oxoid CN0035). Identification of Campylobacter was by characteristic appearance on the culture medium (moist, creamy-grey and flat-spreading), gram stain, oxidase test, catalase reaction hydrolysis of hippurate and indoxyl acetate and confirmed by multiplex PCR (mPCR).
The reference standard strain C. jejuni (LMG 13646) was used for quality control [32].

Molecular Identification of Campylobacter
Multiplex PCR (mPCR) was used for confirmation of the suspected Campylo-  [34]. For genus identification of Campylobacter, 16S rRNA region was amplified using C412F and C1228R primers. Species-specific identification of C. jejuni and C. coli was done by targeting the hipO gene which codes for hippuricase enzyme with a 344 bp, and the aspK gene coding for aspartokinase with a 500 bp, respectively ( Table 1). The PCR products were separated and identified by gel electrophoresis using 1.8% agarose containing 0.1 μg/ml ethidium bromide.

Data Management and Analysis
The data were entered and analyzed using SPSS (IBM Corp., Armonk, NY, USA) version 21. Binary logistic regression was used to model predictors of Campylobacter infection adjusting for factors identified to be significant at P < 0.05 in bivariate analysis. Backward conditional logistic regression was used with removal of predictors at P < 0.05. Odds ratios (OR) and 95% confidence intervals (CI) were calculated to compare association between detection of Campylobacter species and predictor variables in the study.

Socio-Demographic Characteristics
A total of 1200 consenting/assenting diarrheal outpatients seeking health care services at the Busia County Hospital were enrolled. The gender distribution was even between male and female (48.1% vs 51.9%) with a median age of 3.2 years (age range 1 month to 70.2 years). Patients aged less than 5 years constituted the largest group (79.6%; 955) and were stratified as <2 years 535 (44.6%), 2 -5 years (442; 36.8%) and the older group > 5 years (223; 18.6%). Majority (62.3%) of the study subjects were from rural setting ( Table 2).

Prevalence of Campylobacter Species
The overall prevalence of the Campylobacter species of interest, C. jejuni (CJ) and C. coli (CC) was 11.6% from which C. jejuni and C. coli had 89.2% and  10.8% respectively. The recovery rate of Campylobacter species was relatively higher in children < 2 years (13.7%) as compared to other age groups and showed an overall general progressive downward trend with age ( Figure 1). However the difference were not significant (P = 0.088) ( Table 2).

Possible Risk Factors Associated with Campylobacter Infection
Of all the potential risk variables associated with household practices investi-  The most common source of drinking water in this area was borehole (49.5%) followed by treated tap water from municipality (28.4%), river (20.1%), pond (1.9%) with a small percentage from the lake (0.1%). Our analysis showed drinking of untreated pond water, an additional significantly risk to children < 5 years (OR 7.43, P < 0.001) together with female gender (OR 8.25, P < 0.001). Likewise, we noted specific risk variables in >5 years as drinking of raw milk (OR 7.14, P = 0.001) and contacts with infected (diarrheal) person(s) from the same household (OR 4.72, P = 0.006) ( Table 3).
Among the clinical symptoms of Campylobacter infection analyzed, fever (77.7%) and vomiting (53.8%) were the main reported clinical signs. Bivariate analysis revealed that crumps (OR 2.23, P < 0.001), chills (OR 2.77, P < 0.001) and fatigue (OR 2.37, P < 0.001) were significantly associated with the infection in <5 years. However, there was no significant (P > 0.05) association found at the multivariable level.
The childhood expanded immunization program take up was high (84.4%) in this area while very few people (0.5%) were under long term medication. The literacy level for the study participants was high with only 9% reporting no formal education. No significant association (P > 0.05) was found between the level of education of the participants and Campylobacter infections.
The analysis of social economic factors and the participant responses were; type of house lived in (permanent 32.3%, semi-permanent 60.9% and thatched 6.8%), lighting system used (paraffin 36.3%, electricity 41.0% and solar 22.7%) and the cooking fuel (firewood 44.2%, charcoal 52.8%, paraffin 2.5%, electricity 0.4% and gas 0.2%) in the households. However, none of the above social economic factors were significantly associated with Campylobacter infection.

Discussion
In this study, the high number of hospital visits from children < 5 years of age (81.4%) suggests diarrhea, which Campylobacter is mostly associated with, was largely a pediatric disease in this region.
Our overall prevalence of Campylobacter infection among diarrheal patients was 11.6%, comprising of 89.2% C. jejuni and 10.8% C. coli. This prevalence was higher compared to reports from studies in Nairobi 3.1% [31] and Kisii 5.8% [35] in Kenya, Tanzania 9.3% [36] and Egypt 9% [20], consistent to those reported from Ethiopia 11.6% [37], and lower than those found in Nigeria 16.5% [38] and Algeria 17.7% [39]. These variations may arise due to differences in the set predictive values of test systems, social and cultural practices, geographical location, population immunity as well as differences in public health standards, intervention strategies, surveillance systems, food safety practices, and the prevalence of Campylobacter in natural reservoirs in different regions [23] [40].
From an African perspective, where Campylobacter is endemic and chicken meat on top feeding list, our reported prevalence (11.6%) in human is considered significant especially in this era of HIV/AIDS. Although this study was li-  ically [43].
Even though there was a general downward trend of prevalence of Campylobacter infection from children < 2 years (13.7%) through 2 -5 years (10.2%) and >5 years (9.4%), this general progressive decline of the infection with age might be due to age-related acquisition of protective immunity as a result of close proximity and frequent interactions with livestock hence frequent exposure to Campylobacter infection [17] [18] [43]. However, other studies have shown a decline of this immunity at an advanced age > 60 years [44].
This study analyzed the potential risk exposures associated with household practices from several exposure variables of social demographic, clinical, economic and behavioral factors considered on the basis of biological plausibility and previous knowledge on the risks associated with Campylobacter infection [13] [20] [21]. Our findings suggest that homestead practices of poultry farming, eating chicken meat, drinking raw milk, breast feeding, use of untreated pond water, contact with infected diarrheal individuals and domestic pets, use of same food cutting board without washing with soap, use of toilets, residential area and gender were significantly associated with human Campylobacter infection in Busia County. Although this study did not comprehensively evaluate all the plausible social economic factors, the few that were analyzed did not yield any significant association with Campylobacter infection.
Majority of the participants presented with fever (77.7%) and vomiting (53.8%) as the main clinical signs. Similarly none of the analyzed clinical signs and symptoms were found to be associated with Campylobacter infection. This is likely because majority of Campylobacter infections in low resource settings are mainly asymptomatic [22]. most of the household livelihoods in Busia County where the study was conducted, depend majorly on livestock farming with chicken being the most commonly kept livestock that offers quick sale and food in an environment where water sanitation and hygiene (WASH) pose a lot of challenges in the livestock sector [29]. Campylobacter enters food chain through poultry regarded as the main source through which transmissions occur and once established in chicken caeca, the number proliferates rapidly to high levels (109 CFU/g) and can remain carriers until slaughter [45] [46]. During this period, high number of Campylobacter are shaded through feces and contaminates the surrounding environment where the birds live and in turn, allow the pathogen transmission to humans through fecal oral rout [47]. Likewise, there has been increasing evidence that contact with farm animals offer significant risk exposures to human Campylobacter infection [13] [48] with poultry being the major source of food-related transmission and reservoir of Campylobacter species to humans [49]. In addition, source-attribution studies have linked zoonotic transmission of human Campylobacter to poultry exposures and consumption of chicken, dairy products and contaminated water [13] [14] [17].
We noted at the same time that household use and drinking of untreated pond water was a significant risk for Campylobacter infection among children < 5 years. This environmental surface water (pond) is served by several contaminated streams from free-range small hold livestock farm systems and can remain stagnant for a long period of time. Consequently such cumulative contamination by Campylobacter species may present potential reservoir and transmitting vehicles for Campylobacter [50] [51] [52] [53]. Studies have shown many possible sources of Campylobacter infection in humans including contaminated water [54]. In Kenya, approximately 17,100 children under 5 years die each year from diarrhea, with 90% of these deaths being attributed to poor water, sanitation and hygiene [55]. According to Busia County Development Plan, 2018-2022, the main source of drinking water is from borehole (46%), other sources including springs, rivers, piped and pond water. Apart from urban treated water supplied by Busia Water and Sewerage Company with a supply capacity of only 12.5%, the larger percentage of water used is untreated. Mothers or caregivers possibly struggle for clean and safe water in this area and a number of them likely end up using untreated pond water to care for children whose immunity may have not developed as much.
Female children < 5 years were more at risk of Campylobacter infection. Similarly, studies on diarrheal correlates associated with enteric bacterial infections in Busia and Muranga Counties in Kenya showed that female between 6 -36 months were twice more likely to be infected with enteric bacterial pathogens than their male counterparts [56] [57]. Likewise in rural Sudan, female children < 5 years were more at risk [58]. Although female subjects were found to be significantly more at risk of Campylobacter infections, other studies have contrasting reports with unclear reasons [59] [60]. Sex specific age differences in immunity may exist but have not been elaborately documented. More investiga-tions are recommended for better understanding on factors such as immunological, demographic and behavioral associated with sex specific differences across the globe. Although we did not directly assess the risk associated with each and every step in handling and preparation of food in household kitchens, assessment of use of same food cutting board without proper wash was picked as a significant risk for children < 2 years. Care givers involved in food preparation for these children were three times (AOR 3.32, P = 0.014) more likely to predispose their children to Campylobacter infection and this is consistent with other findings [61]. In Kenya, complementary feeding is introduced after 6 months though some mothers introduce weaning earlier. It is possible that there might have been lack of personal hygiene and inadequate cleansing of food surfaces by the mothers or caregivers and subsequent cross-contamination of Campylobacter during food preparation from raw food. Similarly, it has been shown that unhygienic food preparation, food storage and subsequent feeding of infants can aggravate transmission of diarrheal pathogens and pose great risk to this age cohort [62]. Otherwise self-contamination and cross transmission of Campylobacter infection may be minimize through personal and kitchen hygiene practices at household level.
Breast feeding and frequent use of toilet had protective effect among children < 2 years. This might be due to transfers of maternal antibodies in breast milk with health befits to the child. Open defecation predisposes water and food to fecal contamination more so when it rains and may increase the risk of Campylobacter transmission. Frequent use of toilets partly takes care of this public health problem and ensures standards necessary to break the vicious cycle of diarrheal pathogens spread from persons infected to the environment hence minimize the transmission of Campylobacter infection and offer protection to these children [39]. In addition, this study found a fairly high literacy level among the participants with 91% having at least acquired primary school education. This may have contributed towards adherence to the recommended WASH practices such as frequent usage of toilets.
Many people keep cats and dogs at home as pets and for security reasons and our study showed this interaction was significantly associated with Campylobacter infection in children of 2 -5 years (OR 5.72, P = 0.016). This might have been due to their playful nature at this age and a likelihood of being physically closer to these pets and more likely than others to contaminate and put their fingers or other items into their mouths which might have aggravated the transmission of the infection. Our finding is similar to previous studies performed in Kenya [63] which showed that the incident of Campylobacter infection was significantly higher in children who had contact with domestic animals and more so in those who keep dogs and cats as pets [64]. On dietary habits reported at household level, consumption of chicken was more frequent (78.2%) of all the foods considered and was significantly associated with Campylobacter infection in children of 2 -5 years. This age is ex-T. T. Ouko et al. plorative, very playful and is common when self-feeding practices occur with subsequent development of immunity due to exposures. When chicken is prepared under unhygienic way or consumed when under-cooked, it can cause Campylobacter infection and is considered the most common cause of diarrheal outbreaks in Europe [65]. Although Campylobacter infection rate was the same among rural and urban dwellers, children between 2 -5 years of age from urban were more protected from the infection 0.47 (0.20 -0.82) P = 0.041 as compared to those from rural. The expected improved environment and water, sanitation and hygiene (WASH) infrastructure in urban areas may have played a role and came with health benefits to these children [66].
Drinking raw milk was the third most used food item at home (70.3%) and was a risk in >5 years. Raw unpasteurized milk has largely been implicated in several Campylobacter infection and outbreaks [67] [68] [69]. Elsewhere, genomic analyses showed that the presence of Campylobacter species in milk was attributed to fecal contamination [70].
We found exposure to household member with diarrhea a risk to those > 5 years of age (AOR 4.72, P = 0.006). The mode of transmission here is thought to be through fecal-oral root from either person-to-person or exposure to a com-

Conclusion
The present study found that diarrhea is mainly a pediatric disease and children < 5 years are more predisposed to homestead related risk activities associated with Campylobacter infection. Campylobacter jejuni was found to be the leading cause of Campylobacter infections in the study area. Ruling out Campylobacter infection in patients with diarrhea would help in proper management more so in children who are most affected, and where HIV prevalence is high as in most developing countries. Personal hygiene practices and proper animal husbandry especially where livestock-human interaction is common is important. Further studies are required on sex specific age difference, social economic factors and the role played by the environment and other domestic animals in the transmission of Campylobacter infection. This would advance knowledge and understanding on source attribution and transmission dynamics for effective control and management of the infection.

Data Availability
The datasets used and analyzed during this study are available from the corresponding author on request.

Ethical Consideration
The study was approved by the Scientific Ethical Review Unit (SERU), Kenya Medical Research Institute (KEMRI protocol No. 3320). Permission was obtained from County Health offices and Hospital administrators. Written informed consent was obtained from the study patients or guardians of minors/study patients. The results were shared with the duty physician for appropriate patient treatment and management. Advances in Microbiology