Q fever is an old and neglected zoonotic disease in Kenya: a systematic review

Q fever is an acute (on occasion chronic) zoonotic disease of great public health importance worldwide. The disease is caused by an obligate gram-negative bacterium; Coxiella burnetii [1]. Coxiella (C). burnetii belongs to the genus Coxiella of the gamma subdivision of Proteobaccteria along with the genera Legionella, Francisella, and Rickettsiella. Unlike the other members of Proteobaccteria, C. burnetii is highly resistant to adverse physical conditions and chemical agents, so it can survive for months and even years in the environment. Its preferred target cells are macrophages located in body tissues and the monocytes circulating in the blood stream [2]. C. burnetii exists in two distinct antigenic forms, the phase I and phase II bacterial variants which can be discriminated by the surface lipopolysaccharide (LPS) composition. This antigenic variation is important for serological diagnosis and pathogenesis. Phase I variants are the highly infectious forms found in naturally infected hosts whereas phase II variants are less infectious and are obtained after serial passages in cell culture systems or embryonated eggs [3].

Domestic animals such as cattle, sheep and goats act as the major reservoirs of C. burnetii which can infect a large variety of animals, humans, birds, and arthropods [4‐7]. Human infection results from inhalation of contaminated aerosols, consumption of contaminated unpasteurized dairy products, direct contact with contaminated milk, urine, feces, or semen of infected animals, and tick bites [8, 9]. Clinical presentation is nonspecific and highly variable ranging from asymptomatic infection (60 %) or self-limiting febrile illness associated with fatigue, headache, general malaise, myalgia, arthralgia, to atypical pneumonia(rapidly progressive courses may occur) and/or hepatitis. Less frequent complications include endocarditis, osteomyelitis and aseptic meningitis. About 1–2 % of acute symptomatic cases may develop chronic disease [10, 11]. Q fever is considered to be an occupational disease of people who have intimate contact with animals or their products such as veterinarians, farmers, abattoir workers, and laboratory workers [12]. There is emerging evidence of C. burnetii as a cause of non-malaria febrile illness and community acquired pneumonia in many developing countries [13‐18], but hospital based diagnosis of Q fever in Kenya is uncommon. The lack of attention is mainly due to scarcity of available data and the perceived low clinical relevance of Q fever in relation to other endemic fevers. As a result, the disease might often be underreported and the disease burden grossly under-estimated [19].

In animals, Q fever is frequently asymptomatic. Sheep and goats may exhibit abortion, stillbirth, premature delivery, and delivery of weak offspring while cattle and camel may develop infertility, metritis, and mastitis [20, 21]. Animal studies have demonstrated that vaginal mucus, feces, and urine are the common shedding route and means for environmental contamination through kidding and effluent mismanagement [22, 23]. Mammals also considerably shed C. burnetii in milk and thus consumption of contaminated unpasteurized milk or dairy products can be a significant source of human infection [24]. Similar to humans, Q fever is under appreciated as cause of animal disease in Kenya possibly leading to persistence of the infection in animal herds, impacting on livestock productivity, and acting as sources of infection for humans [15].

The past surveys show strong evidence of Q fever exposure in animals and humans in all regions in which studies were conducted and the seroprevalence rates appear high. This is in contrast to the scarce literature available (Table 1). Though the first Q fever case was reported 63 years ago in Kenya [26, 35], only limited high quality epidemiologic studies utilizing randomly sampled human-animal populations for Coxiella have been conducted. Available studies were mainly retrospective using archived samples either submitted to reference laboratories or collected in hospitals for other diseases diagnosis resulting in sampling bias. We found no descriptions of official national disease surveillance and reporting or control programs in the published literatures. This knowledge gap highlights the need for more representative epidemiological studies in the country to elucidate the exact disease burden and pathogen transmission dynamics involved. The high seroprevalence of human antibodies against Coxiella antigens in the previous studies (3–35.8 %) suggests that cases are often misdiagnosed and/or unrecognized potentially leading to wrong patient management. This is a great challenge also in many sub-Saharan Africa countries where other tropical fevers such as malaria, typhoid, rotavirus and pneumonia are endemic. These infections present with similar symptoms making their diagnosis difficult leading to systematic under-reporting in the human health system [40, 41]. Studies in neighboring countries have reported comparable Coxiella seropositivity in humans, i.e., in Tanzania (3.9–20.3 %) [13, 19, 42], in Ethiopia (6.5 %) [43], and in Sudan (10 %) [44].

Q fever is widely classified as an occupational disease for those who have close contact with animals or their products. Lack of appropriate personal protective clothing, present or past residing in a farm are reported to increase ones odds for Coxiella seropositivity [9, 12]. Farming, alcohol intoxication status of abattoir workers and lack of appropriate protective clothing during work were associated with seropositivity in Kenya [33, 34]. However, high quality data related to Q fever awareness or underlying social-economic or cultural factors predisposing humans to C burnetii infection are lacking. No studies were found on populations considered at high risk of infection such as veterinarians, retail butchers, pastoralist communities, and mixed livestock-crop production farmers. A study in Egypt on high risk groups reported high seropositivity (16 and 22 %) in populations living in close contact with animals and those living in rural establishment [45]. In contrast, a study in Chad found low seropositivity (1 %) in similar high risk groups [46]. A study in Tanzania demonstrated an association between Q fever onset and dry season [18]. Our review did not identify investigations on how the pathogen cycles are potentially embedded in livestock production and management systems in the economically and ecologically heterogeneous regions of Kenya.

We identified no study in human populations from northern eastern and upper eastern regions of Kenya but the available animal reports from similar populations in north rift regions revealed high exposure rates of 51.1 % and 20 % in cattle and camel, respectively (Fig. 3) [29, 35]. Epidemiological studies are needed in these high risk regions of Kenya because the communities in these regions practice mixed livestock ranching-and-wildlife conservancy systems and predominantly individual livestock-based pastoralism with high rates of livestock-wildlife interaction. Seroprevalence of C. burnetii of 7.4–51.1 % in cattle, 20–46 % in goats, 6.7–20 % in sheep, 20 and 46 % in camel and 13 % in rodents have been reported [18, 29, 35, 37, 38]. We found no distinct variation in C. burnetii infections among the animal species and the different administrative provinces. However, there were seemingly higher infections (28.2–57.1 %) reported in livestock from arid and semi-arid ecoregions of Rift Valley, Coast, Western and Eastern provinces when compared to Central and Nairobi. Though the precise reason for this variation is not clear, it is possible that specific local variation in prevalence exist based on the differences in animal husbandry practices, human and livestock density patterns and the differences in livestock infection in each of the ecoregions. Therefore, if no appropriate control strategies are implemented, the infected animals may continue to serve as reservoirs and source of infection to the uninfected animals and humans. Moreover, data on other common domestic animals in Kenya such as pigs, donkeys, and cats are not available. Therefore, the role of these domestic animals or ruminant wildlife animals as reservoirs and transmission of Coxiella need to be investigated. Surveys in the neighboring countries have reported Coxiella seropositivity with similar ranges in domestic animals, i.e., in Tanzania (13.3 % cattle, 13.6 % goats, 17.1 % sheep) [42], Ethiopia (31.6 % cattle, 54.2 % goats, 90 % camel) [47] and Sudan (24 % goats, 40.4 % cattle, 53 % goats, 62.5 % sheep [48, 49]. This epidemiologic status may be linked to the unregulated cross-border livestock movements through trade routes and nomadic movements between these countries in search of pasture and watering points that may allow the entry and spread of infected herds. These findings call for further research to elucidate the epidemiology of Q fever in linked animal and human populations in large representative surveys (within this african region) to enable assess the correlations of the disease incidence with agricultural production systems, geographical and climatic conditions and the animal husbandly practices.

Cell culture of Coxiella in a variety of culture media is considered as the gold standard for diagnosis for Q fever. However, the process is usually time consuming and difficult. Moreover, Coxiella is a highly hazardous pathogen whose isolation requires biosafety level 3 (BSL3) laboratory facilities because of its high infectivity potential [9]. As a result, this requires huge investments in terms of specialized skills and equipment. These are not affordable in many developing countries such as Kenya and thus previous epidemiological studies based on bacterial isolation have not been feasible. Recently, a cell-free laboratory culture medium whose composition mimics that of the host cells phagolysosome has been developed and it is undergoing evaluation and validation [2]. This discovery may therefore permit culture based epidemiological studies in resource limited settings such as Kenya in future. Detection of C. burnetii DNA in various samples by a range of Polymerase Chain Reaction (PCR) assays is progressively becoming available and is considered useful especially in the acute clinical stages of the illness when seroconversion is not sufficient to be detected by serological methods [50]. Similarly, the cost of the respective equipment and technical expertise makes it difficult for adoption for routine hospital based testing or to enable seroepidemiological research in many developing countries. Despite the availability of these diagnostic facilities in specialized and research laboratories in Kenya, Q fever diagnosis is still not considered during routine human and animals’ disease diagnosis systems. This can be attributed to the fact that these facilities are located in major towns which are not readily accessible to the regional health care facilities. Efforts to establish a regional laboratory in at least each county with facilities for performing reliable serological assay such as ELISA or serum based real-time qPCR are highly recommended. This approach will shorten the time needed to gain a sensitive diagnosis in patients presenting with fever of unknown origin and enable appropriate management of the patients.

As shown for other countries, controlling of the disease in animals contributes significantly to the decreased incidence in humans [51]. Vaccination of animals against Q fever could therefore present an effective control method. Until now, several vaccines have been developed consisting of inactivated whole phase 1 bacteria. However, many were shown to be nonprotective and presented several limitations such as short term protection and the inability to distinguish between vaccinated and naturally infected animals. Currently, prevention of Q fever in some European countries includes animal vaccination with the non-fully licensed inactivated phase I vaccine, Coxevac (CEVA-Santé Animale, France), after a focus of Q fever has been determined [9]. Several different vaccines have been developed for use in humans. These are mainly composed of a live attenuated strain or subunit vaccines [2, 52]. Despite demonstrating antigenicity and immunogenicity in many trials, humans develop severe reactions to the vaccinations making many of these vaccines effective to only individuals that have no previous exposures. This necessitates extensive screening before administration to having a direct impact on the final cost of the vaccine [53]. Presently, the considered effective vaccine is a formalin killed whole-cell vaccine (Qvax, CSL Limited, Australia) that is licensed for use in Australia, especially in high risk occupational groups. The availability of the vaccines present a great challenge in most of the developing countries and is currently not performed in Kenya.