Seroepidemiological Studies of Crimean-Congo Hemorrhagic Fever Virus in Domestic and Wild Animals.

Crimean-Congo hemorrhagic fever (CCHF) is a widely distributed, tick-borne viral disease. Humans are the only species known to develop illness after CCHF virus (CCHFV) infection, characterized by a nonspecific febrile illness that can progress to severe, often fatal, hemorrhagic disease. A variety of animals may serve as asymptomatic reservoirs of CCHFV in an endemic cycle of transmission. Seroepidemiological studies have been instrumental in elucidating CCHFV reservoirs and in determining endemic foci of viral transmission. Herein, we review over 50 years of CCHFV seroepidemiological studies in domestic and wild animals. This review highlights the role of livestock in the maintenance and transmission of CCHFV, and provides a detailed summary of seroepidemiological studies of wild animal species, reflecting their relative roles in CCHFV ecology.


Introduction
Crimean-Congo hemorrhagic fever virus (CCHFV), a nairovirus of the Bunyaviridae family, is the causative agent of a severe human hemorrhagic fever disease characterized by fever, weakness, myalgia, and hemorrhagic signs [1]. Clinical disease is restricted to humans and is fatal in 3%-30% of cases. Crimean-Congo hemorrhagic fever (CCHF) has been described over a wide geographic area including Asia, Africa, and Europe. The natural vector and reservoir has been identified as Hyalomma spp. ticks, and the distribution of human cases closely mirrors vector distribution. CCHFV is transmitted to humans by the bite of an infected tick, contact with patients during the acute phase of illness, or by contact with blood or tissues of viremic animals. Early diagnosis is critical for patient support and for preventing spread of infection through well-documented human-to-human transmission [2]. Ribavirin has been used extensively as an antiviral treatment, but remains controversial [3,4].
Several groups have published reports of detailed serosurveys conducted recently in various countries, including Albania, Iran, Sudan, and India. However, numerous studies investigating serological evidence of CCHFV in animal species were performed decades ago, are difficult to obtain, and are often published in non-English languages. Animal serosurvey data have been examined and discussed in CCHFV reviews [1,6], but no literature currently exists cohesively presenting current and past reports of the presence or absence of CCHFV antibodies in domestic and wild animals. Virus emergence and reemergence continue to be key topics of national and international health security. As with other hemorrhagic fever viruses, the potential introduction of CCHFV into new geographic areas [29][30][31] should be considered and requires appropriate knowledge of virus ecology, transmission dynamics, and competent reservoir hosts and vectors.
Herein, we provide a detailed summary of the extensive seroepidemiological CCHFV studies performed internationally in both domestic and wild animals (Fig 1). This report serves as an important resource in discussion of the role of animals in CCHFV maintenance and transmission to humans. The information provided specifically aids in understanding the global impact of CCHFV and clarifying the roles of domestic and wild animals in putative expansion of CCHFV endemic regions.

Domestic Animals
Seroepidemiological studies in endemic regions indicate that various domestic and peridomestic animals could be asymptomatically infected with CCHFV. Detection of CCHFV antibodies in domestic animals has been important in providing initial evidence of circulating virus and in localizing CCHFV foci and increased risk for human infection [6,32,33]. A wide spectrum of domestic animal species has been investigated internationally, including cattle, sheep, goats, horses, pigs, dogs, and chickens (Table 1). Other domestic species investigated include buffalo, camels, and ostriches. Examples of high seroprevalence in domestic animals include 79.1% seropositive cattle (Afghanistan) [34], 75.0% sheep (Afghanistan) [34], 66.0% goats (Turkey) [10], 58.8% horses (Iraq) [35], and 39.5% donkeys (Tajikistan) [36]. High seroprevalence has also been reported in camels; the highest (excluding the 1/1 animal found positive in Pakistan) percentage of seropositive camels was reported in Kenya at 26% (n = 499). The largest reported sample size of a single species comprised almost 9,000 cattle tested in South Africa [37]. The role of cattle, sheep, and other large vertebrates in CCHFV ecology is reflected in the relative levels of species-specific CCHFV antibody prevalence reported internationally (Fig 2). Among studies that indicate sample size, cattle are the most often studied (75 studies), followed by sheep (49 studies) and goats (33 studies). Data on cattle and sheep have also been reported from the largest number of countries (34 and 25, respectively) ( Table 1). Reports of other species are more limited; seroprevalence in domestic dogs, for example, was only reported in one study based on samples obtained in South Africa and Zimbabwe [13].
Domestic animal species are often implicated in CCHFV transmission when human CCHF cases are detected. Sheep have been recognized as very important CCHFV reservoirs in certain endemic regions, and have been epidemiologically linked to human cases on several occasions [64,79,84,85]. In Uzbekistan, three CCHF cases were described in persons involved in the handling of tissue from a cow [86]. Similarly, the first patient in an epizootic of CCHFV in Mauritania became ill shortly after butchering a goat [78]. As such, increased CCHFV IgG seropositivity in livestock often parallels reports of CCHF cases in humans with exposure to livestock (e.g., slaughterers, butchers, and farmers), particularly in those who handle blood and organs from infected livestock [34,[87][88][89][90][91][92]. Conversely, negative seroprevalence results in domestic animal samples reflect either low-level transmission or the absence of CCHFV in those geographic areas. Thus, no evidence of seroprevalence in domestic animals was found in samples from Germany, Italy, the Netherlands, Australia, or New Zealand, all countries with no CCHFV cases reported to date [57].
The tick-vertebrate-tick cycle of CCHFV maintenance is reflected in relative tick abundance and associated animal seroprevalence. Cattle heavily infested with ticks were more likely to be CCHFV seropositive [26,75], and vector control to reduce the tick burden was associated with decreased seroprevalence [75]. Cattle are noted as the most sensitive indicator of low-level CCHFV circulation because they tend to be highly infested with Hyalomma spp. ticks, the numbers of which can be ten times higher than those found on small ruminants [93]. In Iran, following detection of human CCHFV cases in Kurdistan Province in 2007, ticks were collected from cattle, sheep, and goats. Of the collected ticks, 5.6% (5/90) were positive by reverse transcription PCR for CCHFV, and four of the five positive ticks were collected from cattle [94]. While there appears to be an association between the presence of infected ticks and detection of seropositive animals [95], viral RNA in attached ticks does not directly indicate seropositivity in host species, and vice versa: infected ticks have been found on seronegative animals and uninfected ticks on seropositive animals. Abiotic variation by season, country, and region is reported in CCHFV seroprevalence studies. Studies in Turkmenistan (then Turkmen Soviet Socialist Republic [SSR]) reported an increase in CCHFV seropositive domestic animal species during the summer season, and found large variations between regions and individual farms (seropositivity range 5.9%-32%) [32]. Geographic variation of CCHFV seroprevalence in domestic animals within a single country has also been reported in several studies [10,51,61,82]. Longitudinal studies in Russia (Rostov Oblast) demonstrated considerable variation when repeated sampling was performed in the same location. These studies reported September as the optimum period for detecting precipitating antibodies in this area, with a notable decrease in seroprevalence in the winterspring period [71]. In support of the recognized endemic transmission cycle of CCHF, variation in seroprevalence is often associated with competent vector distribution, host preference of competent tick vectors, and tick load on a particular animal species. Anti-CCHFV antibody prevalence is highest in biotopes where Hyalomma spp. ticks often predominate. Sustained endemic transmission is found only where Hyalomma spp. ticks are present, and epizootic transmission occurs during periods of increased abundance of these ticks [96]. In the hyperendemic CCHFV region in Turkey, the overall tick infestation rate of livestock was 61.2%; 63.1% of cattle and 56.9% of sheep were infested with one or more tick. The dominant species infesting both cattle and sheep was Hyalomma marginatum [97].
A subset of biotic factors determining domestic animal CCHFV seroprevalence were investigated in Senegalese sheep by Wilson et al., who reported that the sex of the animal did not affect antibody prevalence [83]. Other factors, including increasing age, are consistently associated with higher seroprevalence in domestic animals [26,27,61,75,82]. Age likely reflects repeated exposure potential, as described by Adam et al., who found that calves started to get infected after the age of two, the age at which they are released to pasture for grazing and, thus, are more likely to be exposed to infected ticks [75]. Breed may also play a role: in Sudan, crossbred cattle were 37 times more likely to be seropositive than endogenous breeds [75]. Further insight into the dynamics of infection in domestic species was provided by a longitudinal serosurvey conducted by Zeller et al. in Senegal from 1989 to 1992 [95]. Investigators collected ticks feeding on two cows and 12 goats, and obtained paired blood samples three times per month. Seropositive animals infested with infected ticks had even higher anti-CCHFV IgG antibody titers than seropositive animals without ticks, supporting the occurrence of reinfection in domestic species. The persistence of anti-CCHFV IgM antibodies in naturally infected animals was found to be 1-2 months [95].
Studies in companion animals are very limited and, thus, difficult to broadly interpret. Antibodies to CCHFV were reported in 6% (n = 1978) of dogs in South Africa and Zimbabwe [13]. In another study, in association with human CCHF cases in Mauritania in 2003, feeding ticks were collected from livestock and dogs. A proportion (five of 56 tested) of Rhipicephalus evertsi evertsi ticks collected from sheep were found to be CCHFV positive by reverse transcription PCR, but none of the five Rhipicephalus sanguineus ticks collected from dogs were positive [78]. While vector competence and host preference may indicate the risk of natural infection and transmission in companion animal species in the absence of serological data, broadly translating vector data to risk of exposure remains complex, as tick data is not always consistent and is influenced by many factors unrelated to the host. For example, CCHFV has been isolated from Rhipicephalus spp. ticks [98]. However, R. sanguineus (brown dog ticks) have been reported as positive or negative for CCHFV depending on the study [44,99]. Additional data on companion animals and associated vector species will aid in more clearly evaluating the role of companion animals in the ecology of CCHFV.

Wild Animals
The seroepidemiological reports of CCHFV in wild animals reviewed herein comprise almost 7,000 samples from over 175 avian, mammalian, and reptilian species (Table 2). Considerable seroprevalence was consistently reported in hares (3%-22%), buffalo (10%-20%), and rhinoceroses (40%-68%). Of the species investigated, those with low reported seroprevalence include elephants (single animal), marmots (no evidence), all non-human primate species (no evidence), and all insectivore rodent species (no evidence). While anti-CCHFV antibodies were not detected in Insectivora rodents, several seropositive hedgehogs (Erinaceus europaeus, Hemiechinus auritus) have been reported, and a substantial tick load of up to 40 larval and nymphal H. marginatum ticks has been described on hedgehog hosts during the peak season of immature tick activity [6,100]. However, the role of hedgehogs in enzootic maintenance appears to be variable by species. H. auritus develop viremia during experimental infection [101] and are considered a natural CCHFV reservoir by serving as a source of CCHFV for feeding ticks. In contrast, in the same study, experimental infection in the European hedgehog (E. europaeus) did not produce detectable viremia, suggesting reduced susceptibility to infection or more efficient viral clearance.
Two reports have found antibodies to CCHFV in representatives of the mammalian order Chiroptera. Using the AGDP test with antigens prepared from CCHFV strains isolated in then-Soviet republics, antibodies were detected in blood sera from two bats in France, from an area bordering with Spain [104]. The species sampled were not specified, and this remains the only report of CCHFV seroprevalence in France. One additional study in northern Iran reported evidence by AGDP in Chiroptera species, in the sera of the greater mouse-eared bat and the common noctule [9]. While these reports appear to be the only evidence of CCHFV infection in bats, recent investigations into bat viruses suggest that there are other species of nairoviruses circulating in bat populations. Using modern sequencing techniques, the first bat nairovirus was identified in French insectivorous bat specimens [107], and a novel nairovirus was isolated from Zambian bats [108].
In reptiles, anti-CCHFV antibodies were detected in one Horsfield's tortoise (Testudo horsfieldii) trapped in early June in Bul'yoni-Bolo winter camp in the Dangara region of Tajikistan [36]. There are several conflicting reports as to the total sample size of the study, ranging from four to 209 tortoises [6,36]; reported seroprevalence is based on the most detailed report provided by T.P. Pak [36]. Other limited investigations of reptile samples did not yield any evidence of antibodies to CCHFV [36,76]. However, a recent report detected CCHFV in Hyalomma aegyptium [109], the tortoise tick, suggesting that tortoises may be similar to certain bird species (discussed below), in which infected ticks are commonly found feeding on the animal, and CCHFV transmission to ticks may occur even in the absence of detectable antibodies in the host.

Birds
Many bird species are important hosts for Hyalomma ticks and can transport ticks over long distances [110,111]. The transport of CCHFV-infected ticks by birds is a current topic of concern regarding regional spread of the virus [29,112]. Historical studies found birds associated with cattle pastures to be important in feeding immature tick stages, and that rooks (Corvus frugilegus) were particularly important; an increase in CCHF cases was associated with increased rook populations [113]. However, CCHFV infection and the presence or absence of an antibody response in avian species remains unclear. The majority of serosurveys of wild avian species report no serological evidence of CCHFV infection in birds, despite investigation of numerous species and substantial sample pools (Table 2). This absence of viremia is      interesting, as some species support large numbers of CCHFV-infected ticks [6,69]. This observation has been supported by experimental infection; the red-billed hornbill (Tockus erythrorhynchus) was found to replicate CCHFV without detectable viremia and was able to infect immature Hyalomma rufipes ticks [114,115]. However, another experimental infection study of mostly ground-feeding birds suggested that anti-CCHFV antibodies may be produced following infection; blue-helmeted guineafowl (Numida meleagris), for example, developed lowlevel viremia followed by a transient antibody response [17]. Studies on Anseriformes and Galliformes species are also conflicting. In pathogenicity studies, experimentally infected domestic chickens were found to be refractory to CCHFV infection [17]. However, a 0.2% CCHFV seroprevalence in chickens and ducks (n = 428) was reported in Kazakhstan [66]. The absence of detectable anti-CCHFV antibodies in birds may reflect limitations in assay sensitivity. Most of the serological surveys on birds in the former USSR were based on the AGDP test [6], and several studies have shown that the AGDP test is less sensitive than the RPHI or IFA tests for detection of CCHFV antibodies [13,17,37]. More recent investigations, however, suggest that past reports accurately reflect the absence of antibody production, and that most species of birds do not appear to develop viremia. An investigation by Shepherd et al. on the sera of 460 birds of 37 species failed to detect antibodies to CCHFV [17]. However, the absence of antibody production is not universal to all bird species. Ostriches appear to be an exception amongst avian species in harboring and possibly transmitting CCHFV to humans. In the above-mentioned studies by Shepherd et al., anti-CCHFV antibodies were found in 22/ 92 (23.9%) ostriches (Struthio camelus). Of note, antibodies were detected in 6/9 (66.6%) ostriches in association with a human CCHF case in a worker who became ill after slaughtering ostriches on a farm in South Africa [17]. Additionally, 1/5 (20%) ostriches tested in association with four CCHF cases in workers from two ostrich farms in Iran were also found to be positive for CCHFV IgG [80]. Experimental infection has shown that viremia in ostriches is very short in duration [116].

CCHFV Isolation from Animals
Experimental studies suggest that many animal species develop a transient viremia, and thus may play a role in transmitting CCHFV to ticks in nature. However, reports of CCHFV isolation from animals are limited. CCHFV has been isolated from a febrile cow in Kenya, cattle and a goat in a Nigerian abattoir, a goat placed as a sentinel for arboviruses in Senegal, European hares in Crimea, and a hedgehog in Nigeria (Table 3). Further supporting serological data, in an extensive study in endemic foci in Russia (Astrakhan Oblast), no virus was isolated from over 350 bird specimens representing 35 species. The paucity of CCHFV isolates from animals likely reflects a relatively brief viremic period and difficulty in identifying infected animals due to absent or mild clinical disease [119][120][121]. The majority of reported CCHFV isolations are from ticks or human case-patients. This is a result of an increased relative likelihood of isolation and, in turn, a preference for tick and human case specimens for isolation attempts. However, inability to isolate CCHFV from vertebrate animals does not necessarily indicate a lack of infection in these animals, and does not rule them out as potential CCHFV hosts capable of spreading disease to humans.

Discussion
A large amount of research investigating the role of animals in transmission and maintenance of CCHFV was performed beginning in the late 1960s and 1970s. This work was instrumental in identifying mammalian species, particularly livestock, as critical in the maintenance of CCHFV and as sources of human exposure. The knowledge gained from these studies has also been important in developing prevention and control strategies such as the use of acaricides on livestock in endemic regions. Recently, numerous studies have provided additional information on known reservoir species and provided country-specific information on animal species with notable roles in CCHFV maintenance.
The reports summarized herein must be considered broadly and examined for trends and not specifics due to several factors. Reported seroprevalence may be biased by sample size, seasonality, and diversity in sampling sites, since if one animal is seropositive, additional positive animals are likely to be found in that location at that time. In addition, these reports used a variety of serological assays. There are caveats to interpretation of individual assay results [12], and direct comparison of results from a variety of assays is confounded by variation in assay sensitivity and specificity. Several groups have performed direct comparisons of the reported serological assays [20,36,69,122]; however, results of the comparisons themselves will vary depending on the conditions of the specific assay and the species investigated. Also, several iterations of the same format of serological tests have been used over the years, making generalized statements about their relative reliability challenging. Comparison of serological techniques for use in animals has been performed for other zoonotic viral hemorrhagic fevers [123]. For CCHFV, the merits and pitfalls of several of the serological assays were reviewed by Hoogstraal [6], who advises that most earlier seroepidemiological results be regarded as suggestive of CCHFV seropositivity but not as positive proof.  Goat Capra spp. Nigeria 1 [117] Senegal (Bandia Forest) 1 [6,18] Hedgehog Hemiechinus auritus Crimea 0/17 [118] Erinaceus albiventris Nigeria 1 [117] Misc. birds Russia (Astrakhan Oblast) 0/360 In [6] doi:10.1371/journal.pntd.0004210.t003 Overall, serological detection methods have improved over time. Technological advances, including the advent of ELISA assays, allow detection of low amounts of infectious virus or of inactivated antigen and antibodies to CCHFV, and have been shown to be more sensitive, specific, rapid, and reproducible than CF, IFA, RPHI, or AGDP [124]. ELISAs are generally considered the preferred method of serological investigation for CCHFV. However, sandwich ELISA techniques cannot be applied successfully to all species [28], necessitating further advances in testing, including a CELISA that was validated during an extensive CCHFV serological survey in South Africa [28]. Of note, species-specific validations of ELISAs have been performed; Qing et al. evaluated a recombinant nucleoprotein-based system for IgG detection in sheep sera [125], and Mertens et al. developed an ELISA for CCHFV IgG antibodies in bovine sera, showing it to have >98% diagnostic sensitivity and specificity [24].
Finally, there is also the potential for cross-reactivity with other related nairoviruses such as Dugbe virus, Nairobi sheep disease, and Qalyub viruses [20,25]. Antibodies to other nairoviruses may exist independently or in conjunction with CCHFV-specific antibodies. Thus, reports of seroprevalence in areas not previously identified to have CCHFV transmission would benefit from additional surveillance, such as tick studies, to help support novel identification of CCHFV foci.
Irrespective of the nuances of serological assay interpretation and incongruity, the data from the studies summarized here, importantly, indicate broad areas with endemic transmission and highlight reservoir species with the highest potential to affect public health. Some species may serve as direct sources of viral transmission (e.g., viremic livestock, ostriches), whereas others aid principally in maintaining high levels of CCHFV endemicity (e.g., hares). These data also highlight species that could present a risk but have not previously been implicated in human cases, such as camels that are replacing cattle use in certain regions due to climate change [126].
With extensive areas of endemic transmission, the issue of CCHFV importation via animal hosts, ticks, or human cases is a critical concern. Importation of livestock was highlighted in a 1994-1995 CCHFV outbreak in the United Arab Emirates; CCHFV sequences from the patients of this outbreak were identical or closely related to those from three Hyalomma spp. ticks obtained from livestock recently imported from Somalia [127]. It is not clear, however, whether the imported animals were infected at the time of importation or more susceptible to infection upon arrival. Williams et al. [46] reported higher seroprevalence in imported sheep and goats than in indigenous animals, which was attributed to increased susceptibility of naïve animals and virus circulation within the quarantine areas. A subset of the sheep sampled was from Western Australia, a region in which no CCHFV-competent vectors have been reported. The majority of imported animals surveyed from Australia had been in Oman for more than 30 days and, although reported as tick-free upon entry, had high levels of Hyalomma spp. infestation at the time of sampling, providing opportunity for CCHFV exposure. Importation of human cases has also occurred. To date, four human cases of CCHF have been imported into a non-endemic country: in 2004, a case was imported into France from Senegal [128]; in 2009, a US soldier entered Germany from Afghanistan; in 2012, an infected person arrived in the United Kingdom from Afghanistan; and in 2014, another came into the UK from Bulgaria. Other unconfirmed reports include a suspected case imported to the UK from Zimbabwe in 1997 and into Germany from Bulgaria in 2001 [129].
CCHFV is widely distributed, circulates in numerous vertebrate species, and can be transmitted to humans in several ways. Serosurveillance of animals will continue to be an essential tool for monitoring levels of endemic transmission and for investigating areas where CCHFV is not known to circulate. The importance of timely assessment of the potential role of domestic and wildlife species in disease introduction and emerging disease response is very important in the case of CCHFV. Our report summarizes data from international studies investigating the presence of antibodies to CCHFV in domestic and wild animals. We provide comprehensive species-specific information and highlight the appropriate literature serving as a critical resource in future discussion of putative importation and extension of known CCHFV endemicity.
those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.