Prevalence and risk factors associated with Ehrlichia infections in smallholder dairy cattle in Nairobi City County, Kenya

Background and Aim: Ehrlichiosis caused by Ehrlichia ruminantium is a tick-borne disease of great economic importance in cattle production worldwide. Despite its economic impact, limited knowledge is available on its epidemiology in Africa, including Kenya. Suspected cases of E. ruminantium infections have been reported in the recent past to the University of Nairobi’s Veterinary Hospital, prompting the need to investigate their possible re-emergence. Therefore, this study was aimed at determining the prevalence of E. ruminantium among smallholder dairy cattle in Nairobi City County and to assess potential risk factors. This knowledge may guide the development of appropriate control strategies of ehrlichiosis, subsequently reducing associated losses. Materials and Methods: A total of 107 smallholder dairy farms from Nairobi City County were recruited for the study. Blood samples were collected from 314 apparently healthy dairy cattle, and Giemsa-stained blood smears were screened under the microscope for Ehrlichia species. A commercial antigen enzyme-linked immunosorbent assay (ELISA) kit was then used to confirm the presence of the infections in serum samples. A pre-tested questionnaire was used to collect data on management practices that may be potential risk factors. A univariate and mixed-effects logistic regression was then used to determine significant risk factors. Results: On microscopy, 79.3% (249/314) of the sampled animals had Ehrlichia-like inclusion bodies in white blood cells, though only 18.6% (95% confidence interval [CI] 14.2-23.0) of these were confirmed to be E. ruminantium on ELISA. A farm-level prevalence of 35.5% (95% CI 27.0-45.3) was reported. Female-headed households (p=0.013), farms in Langata region (p=0.027), cleaning of cowsheds fortnightly (p=0.019), and roofing of cowshed (p=0.022) were factors significantly associated with E. ruminantium infections. Conclusion: There is a relatively high prevalence of E. ruminantium infections in apparently healthy cattle in smallholder dairy farms in this area, warranting control measures. It is critical to improve animal welfare-related factors, such as cowshed cleaning and roofing, as well as the strategic location of farms, especially, since reservoirs may reduce infection levels in the farms, in relation to wildlife. However, since Ehrlichia-like inclusion bodies other than those of E. ruminantium were observed in this study, there is a need to investigate further these factors and the possibility of other Ehrlichia species infecting cattle in the study area.


Introduction
Ehrlichiosis is a broad term referring to tick-borne diseases (TBDs) of multiple animal species caused by obligate intracellular organisms of the white blood cells [1]. Ehrlichia ruminantium is the main Ehrlichia species infecting cattle in Africa and the Caribbean Islands [2]. It is transmitted by infected ticks in the genus Amblyomma [2,3]. Despite E. ruminantium being among the most important TBD affecting millions of cattle in sub-Saharan Africa and being an OIE listed disease [4][5][6][7], there is limited information on its epidemiology in Africa, including Kenya [8]. Instead, other TBDs, such as East Coast Fever, anaplasmosis, and babesiosis, have been considered the most important for cattle throughout Eastern Africa, including Kenya [9,10]. This may be a consequence of certain reports of the low prevalence of E. ruminantium in some parts of Africa and even Kenya, including 0.6% in Ethiopia [11], 0.4% in Kenya [12], and 1.1% and 1.7% in Nigeria and Uganda, respectively [13,14]. Due to the high demand for milk, market forces and the convenience of urban areas, there has been a constant rise in dairy cattle keeping in urban areas of Kenya [15]. These smallholder dairy production systems in urban and peri-urban areas experience unique challenges where their substandard animal husbandry practices compromise animal welfare, thereby predisposing the cattle to stressful conditions and, as a result, various diseases [16].
TBDs are the most important diseases affecting these smallholder dairy farms, mainly due to dynamic tick vectors, increasing the susceptibility of the exotic breeds kept in the farms [17]. E. ruminantium infections cause severe economic losses in Africa, where approximately 150 million animals are at risk of infection [7], thereby negatively affecting livelihoods that depend on cattle. The estimation of these losses in endemic areas is complicated by the fact that farmers do not provide regular reports, definitive diagnoses are hardly delivered [7], and infections often coincide with other TBDs such as anaplasmosis and East Coast Fever [2]. Despite this, the few economic studies that have been carried out have pointed to quite substantial economic losses. In Zimbabwe, losses of USD 5.6 million/year have been reported, with acaricide costs accounting for 76% and milk losses accounting for 18% of the total cost [18]. In a 3-year follow-up study in Ethiopia, losses from mortalities, cost of acaricides and antibiotics, and losses in milk and meat have resulted in total costs of USD 7884.67 million [8]. In Tanzania, economic losses of USD 22.6 million/year have been recorded, with cattle mortality accounting for 8.8 million USD [19].
Suspected cases of ehrlichiosis have been reported since 2014 at the University of Nairobi Veterinary Hospital, which mainly serves Nairobi City County and its peri-urban areas. These cases present with non-specific clinical signs such as inappetence, unthriftiness and, at times, death. When they are screened by a blood smear microscopy, Ehrlichialike inclusion bodies have been shown in the white blood cells of affected animals. The difficulty in diagnosing E. ruminantium infections due to their non-pathognomonic clinical signs has been emphasized by Allsopp and has resulted in the underestimation of economic losses, especially in endemic areas including Kenya [2].
This study was therefore aimed at determining whether the inclusion bodies observed in blood samples collected from smallholder dairy farms within Nairobi City County were those of E. ruminantium, and to assess potential risk factors, therefore warranting the development of appropriate control measures in these farms.

Ethical approval
The study was approved by Biosecurity, Animal Use and Ethics Committee, Faculty of Veterinary Medicine, University of Nairobi (FVM BAUEC/2016/122). Experiments were carried out in accordance with the guidelines laid down by the International Animal Ethics Committee or Institutional Ethics Committee and in accordance with local laws and regulations.

Informed consent
The study was discussed with owners of the animals who were to be recruited and thereafter they signed written consent. They were free to make the decision to join the study or withdraw.

Study area and study population
The study was conducted in Nairobi City County's Nairobi City, which is the capital city of Kenya and is located at 1.28333 latitude and 36.81667 longitude and 1795 m above sea level. The county has 17 sub-counties, the highest number of administrative units in a single county in Kenya. The city has a population of slightly over 3 million people. In the peri-urban areas of the city, there are the smallholder dairy production systems established to meet the high demand for the milk of the city. The previously reported incidence of TBD in these peri-urban areas of Nairobi was 7.8% [10].

Study design
Nairobi City County was divided into four quadrants centered around the city's central business district ( Figure-1). In each of the quadrants, purposive sampling was used to identify the sub-county with the highest cattle population, which was included in the study. The sub-counties identified in each quadrant were Dagoretti, Langata, Ruai, and Westlands. Transect walks were then used to identify smallholder dairy farms in the sub-counties. Smallholder dairy farms were described as those with between 2 and 10 dairy cattle [20]. In each transect the tenth; dairy unit was included in the study depending on the owner's willingness to participate. Otherwise, if they refused to give consent, the next dairy unit was recruited. Animals were grouped into age groups (calves <12 months, yearlings 12 -≤24 months, and adults >24 months). A maximum of two animals in each age group was included in the study. Where there were more than two animals in a specified age group, two were randomly selected. A pre-tested questionnaire was administered to the principal farmer or stockman who spent the most time with the cattle to collect farm-level management factors that were thought to be associated with Ehrlichia infections. Using an expected prevalence of E. ruminantium of 4.5% [21] and sample size calculation formulae according to Charan and Biswas [22], the minimum target sample size was found to be 66 cattle.

Data and sample collection
Data on the various farm-level factors, especially management practices (feeding system, source of fodder, cowshed cleaning, tick control, nature of housing, and introduction of new animal) and animal level factors (age, sex, breed, and lactation status) were collected using the pre-tested questionnaire. Blood was collected from the coccygeal vein and put in vacutainers with clot activator and allowed to stand for 2 h. The serum was then decanted in a labeled Eppendorf tube and centrifuged for 10 min at 3000× g, then decanted into the final Eppendorf tube and stored at −20°C until further analysis. A blood smear was immediately prepared at the farm, fixed using absolute ethanol, and later Giemsa-stained.

Laboratory analysis
Microscopy was done using a light microscope at 1000× oil emersion on the Giemsa-stained smears where the white blood cells (neutrophils, lymphocytes, and monocytes) were assessed for any cytoplasmic inclusion bodies. Antigen enzyme-linked immunosorbent assay (ELISA) for E. ruminantium was done on the harvested serum using commercial kits BIOS microwell ELISA diagnostic systems, USA, following the manufacturer's instructions.

Statistical analysis
Questionnaire data were input into Excel 2016 before it was exported to Stata 15.0 for analysis. Descriptive statistics expressed as proportions were computed for the farm and animal level factors in the study area. Univariate logistic regression was used to test for associations between the various possible risk factors as explanatory variables and E. ruminantium-positive outcome on ELISA. Factors with p≤10% were analyzed using mixed-effects logistic regression model followed by step-wise elimination such that only factors with p<5% were left in the final model and were identified as being statistically associated with E. ruminantium infections.

Microscopic examination
A total of 314 blood samples were collected from cattle in 107 smallholder dairy farms from the four selected regions in Nairobi City County. Inclusion bodies were observed in white blood cells in 79.3%  (Table-1).
Out of the 249 samples that were positive on microscopic analysis, 145 samples that had clean blood smears were scrutinized to assess the specific white blood cells that had Ehrlichia-like inclusion bodies. Of all the cells that were examined, 49.7% (72/145) had Ehrlichia-like inclusion bodies in monocytic cells, while only 4.8% (7/145) were in granulocytic cells. The other 45.5% (66/145) of the samples had inclusion bodies in both granulocytic and monocytic cells. Among the monocytic cells, the lymphocytes were most commonly affected, representing 37.2% of the cells (54/145) followed by monocytes, representing 12.4% (18/145) of the cells. Figure-2 shows the Ehrlichia-like inclusion bodies as they appear in different white blood cells.

Determination of E. ruminantium infections in smallholder dairy farms in Nairobi City County using ELISA
Antigen ELISAs were carried to confirm whether the Ehrlichia-like inclusion bodies were those of E. ruminantium or not. The ELISA was done on 296 of the samples that were examined by microscopy, of which 18.6% (55/296) (95% CI [14.2-23.0])       The majority of animals sampled were adults, representing 48.3% (143/296), and accounted for the highest age group with E. ruminantium infections, representing 10.1% (30/296). The main breed kept in the study farms was Friesian 61.8% (183/296), and, being dairy farms, the majority of animals were females 90.5% (268/296).

Analysis of risk factors associated with E. ruminantium infections among 296 dairy cattle across 107 smallholder dairy farms in Nairobi City County
On univariate analysis, factors found to be statistically significant, at p≤0.1, were the gender of household head, region, farming systems, cleaning of cowshed, type of cowshed floor, presence of cowshed roof, and breed of the animal (Table-3). A mixed-effects logistic regression analysis revealed that the factors significantly associated with seropositivity were farms in which the household head was female (p=0.013), farms from the Langata region (p=0.027), cleaning of the cowshed every fortnight (p=0.019), and farms in which the cowsheds had roofs (p=0.022) (  ruminantium infections than male-headed households. Farms in the Langata Regions were 2.9 times more likely to be infected with E. ruminantium than farms in the Dagoretti region, while farms that cleaned their cowsheds fortnightly were 3.6 times more likely to be infected with E. ruminantium that those that cleaned their cowsheds on a daily basis. Farms in which cowsheds had no roofs were 4.7 times more likely to be infected by E. ruminantium than those which had cowshed roofs.

Discussion
In this study, Ehrlichia-like inclusion bodies were observed in the cytoplasm of various white blood cells. Although microscopic examination of endothelial cells from the brain is the common method of diagnosing E. ruminantium, the inclusion bodies of the parasite can still be observed in the neutrophils on stained blood smears [1,23]. In this study, inclusion bodies were also observed in neutrophils as well as in other white cells such as lymphocytes and monocytes. Some of the samples that revealed inclusion bodies on microscopy were negative on ELISA, possibly indicating the presence of other species of Ehrlichia in the cattle since they all parasitize white blood cells [24]. Despite microscopy being cheap and easily available, the low sensitivity and occasional misdiagnosis due to the presence of artifacts in the smears greatly limit  its use [25]. These limitations may explain the high number of Ehrlichia-like inclusion bodies observed in this study. The current study used antigen ELISA to detect E. ruminantium in an attempt to overcome the limitations linked to cross-reactivity with other Ehrlichia species [26]. E. ruminantium was detected in apparently healthy cattle, an observation similar to that of Matos et al. [27] but inconsistent with Kelly et al. [28], who points out that E. ruminantium always presents with clinical disease. This may suggest that non-pathogenic strains of the parasite may be in circulation, or those carrier animals have the parasites, thereby posing a threat to susceptible animals [2].
The overall prevalence of E. ruminantium in cattle reported in this study was 18.6%, similar to the 15% reported in Mozambique [27], though lower than 50% reported by Swai et al. [29] in Tanzania and 33% in Zimbabwe [26] and higher than 4.5% in Uganda [21] and 4.1% in Ethiopia [25]. This wide variability may suggest differences in factors related to cattle management across various areas. The prevalence recorded in the present study is sufficiently high; however, to warrant the implementation of appropriate control strategies since there may be a risk of clinical disease if susceptible animals are present [29]. The herd prevalence was 35.5%, at nearly twice the animal prevalence, possibly indicating that E. ruminantium infection in this area represents a herd health concern rather than an individual animal's problem [29].
Management practices such as acaricide application and frequency of dipping had previously been reported to be risk factors for E. ruminantium [29,30]. However, these factors were not found to be significant risk factors in the current study. Despite the farmers using acaricide for tick control, therefore ehrlichiosis, its mishandled such as through wrong dilution and application procedures may have led to possible development of ticks' resistance [2,5]. In addition, most of the farms outsourced hay to feed their dairy cattle mainly because they own small land sizes, therefore inadequate fodder for the animals, a common feature of smallholder dairy production units in urban and peri-urban areas in Kenya [16]. The ticks can be carried from the grasses to the stall-fed animals, leading to possible infections.
Farms that cleaned the cowshed fortnightly had increased odds of E. ruminantium infections than those that cleaned them daily. Infrequent removal of slurry from cowsheds has been shown to compromise animal welfare in smallholder dairy units, especially because the housing designs are usually poor, thereby forcing the animals to lie on the slurry for long hours at a time [16]. Poor animal welfare, on the other hand, increases the stress of the affected animals, greatly compromising their immunity and predisposing them to other diseases [31]. This may explain the high number of infected animals in this study. Roofing is part of animal housing that provides protection from extreme weather [32]. Therefore, without roofs in the cowsheds, stressors such as excess cold or heat compromise animals' immunity, possibly explaining the increased rate of E. ruminantium infections in farms with cowsheds without roofs [33].
The farms from the Langata region were more likely to have animals infected with E. ruminantium compared to other regions, possibly because this region borders Nairobi National Park and its wild animals, such as the African buffalo, black wildebeest, and eland, since they have been known to be reservoirs of ehrlichiosis, are associated with increased risk of infection [34].
Farms in which the household head is female were more likely to have animals infected with E. ruminantium than those in which the household head is male. These findings are similar to those of Sungirai et al. [5] who noted that women are less educated in the management of dairy enterprises, thereby increasing their risk of disease. Moreover, Tola et al. [35] also noted that women undertake the majority of the labor in dairy enterprises, minimizing their time for decision making, such as with regards to disease management.
A limitation of the current study was that sub-counties were not randomly selected since this is an urban-set up; therefore, dairy farms were only concentrated in certain sub-counties. However, the farms included in the study were well representative of Kenyan urban smallholder dairy farms; keeping 2-10 animals and practicing sub-optimal animal husbandry as described by Nguhiu-Mwangi et al. [16].

Conclusion
There is a relatively high prevalence of E. ruminantium, both at the animal level (18.6%) and at the herd-level (35.5%), in smallholder dairy production systems in the study region, warranting control measures to be put in place. Farm-level factors that predispose to E. ruminantium infections were identified as infrequent cowshed cleaning, lack of roofing, regions in close proximity to wildlife and female household heads. Moreover, there is a need to further characterize the current strains of E. ruminantium in circulation among dairy cattle in this area, especially to determine their virulence since, as reported elsewhere; some may have the zoonotic potential [36]. Increasing the frequency of cowshed cleaning to remove the slurry and roofing cowsheds will greatly improve animal welfare, subsequently boosting their immunity and controlling E. ruminantium infections.

Authors' Contributions
SGP collected, analyzed the samples, and wrote the manuscript. DWG collected samples and interpreted the data. NM and CMM analyzed and interpreted the data. All the authors read and approved the final manuscript.