1. Introduction
Livestock production in sub-Saharan Africa (SSA) is hampered by several factors including feed and water availability and disease occurrence. These factors are likely to be exacerbated by climate change, particularly in arid and semi-arid ecosystems [
1,
2]. Livestock diseases pose much higher risks in SSA than in any other region because of the tropical climate, often inadequate veterinary services, and uncontrolled animal movement [
3]. In SSA, diseases are responsible for mortality rates of about 7% and 21% in adult cattle and calves per year, respectively [
4].
Rift Valley fever (RVF), caused by the Rift Valley fever virus (RVFV) (genus
Phlebovirus: family
Phenuiviridae), is a hemorrhagic disease affecting both livestock and humans [
5]. It is transmitted to livestock and humans through infected mosquito bites. Other transmission routes include direct contact with shed abortus and infectious fluids, as well as the consumption of raw milk and blood from infected animals [
6]. The clinical signs in cattle include fever, abortion, loss of appetite, and decreased milk production, while in humans, a non-specific febrile illness is common [
6]. The disease is endemic in SSA and several outbreaks have been reported in Kenya. In the last major outbreak of 2006/2007, Baringo County in the northwest of the country was one of the heavily affected regions, with 88 human and 36 livestock cases confirmed [
7]. While the RVF outbreak dynamics and associated losses have been clearly documented [
8], very limited information is available on its inter-epidemic (IEP) transmission and effects on livestock health. Currently, vertical transmission of the virus in
Aedes spp. mosquitoes, horizontal transmission in susceptible livestock/wildlife, intermediate rainfall, and recurrent introduction of the virus from ‘hotspot’ regions are some of the proposed factors responsible for the maintenance of RVF during IEPs [
9].
Bluetongue (BT) and epizootic hemorrhagic disease (EHD) are caused by the bluetongue virus (BTV) and epizootic hemorrhagic disease virus (EHDV), respectively, both orbiviruses from the family
Sedoreoviridae (formerly
Reoviridae) [
10]. Currently, 29 serotypes of BTV and 11 other atypical groupings circulate worldwide [
11,
12,
13,
14]. Unlike RVFV, BTV and EHDV are both transmitted by biting midges, primarily
Culicoides imicola Kieffer (Dipt.: Ceratopogonidae) in Africa [
15]. BT is endemic in Africa, where it affects domestic and wild ruminants; however, outbreaks are usually reported in exotic sheep breeds. The disease in sheep is characterized by fever, facial edema, lacrimation, and oral/nasal hemorrhage [
16]. In susceptible flocks, morbidity can be as high as 100% while mortality rates range from 2–30% [
17]. In cattle, BT is usually subclinical, hence it can circulate undetected; however, some serotypes such as BTV-8 have produced clinical disease in cattle during outbreaks in northwestern Europe [
18]. While BT outbreaks usually have a more devastating effect [
19], the omnipresent subclinical disease is also associated with loss of body weight, a decrease in milk production and reproductive performance, and livestock trade restrictions for infected herds [
20]. A longitudinal study in western Kenya demonstrated that by 1 year of age, 94% of calves had been exposed to BTV and had circulating antibodies, indicating a high prevalence of the virus [
21].
EHD was first described in New Jersey in 1955 in white-tailed deer, where it shows an overt clinical syndrome characterized by per-acute death, fever, oral ulceration, edema, and lameness [
22]. The disease is usually subclinical in cattle, with clinical signs such as fever, inappetence, decrease in milk production, and nasal erosions. For example, in Israel, it was estimated that an EHDV seroprevalence of >80% is associated with a 2% loss in the annual production of dairy cattle [
23]. Seven serotypes of EHDV circulate in several countries in Africa, Asia, the Middle East, and North and South America [
24]. Outbreaks have been reported in Turkey [
24], Israel [
23], Spain [
24], and Tunisia [
25].
Baringo County lies in the postulated BTV enzootic area and is also an arbovirus hotspot [
26]. This is highlighted by previous RVFV outbreaks and the circulation of other zoonotic arboviruses [
7,
27,
28,
29]. Additionally, occasional flooding events from the Baringo and Bogoria Lakes and the high annual temperatures create a conducive environment for the proliferation of mosquito and culicid vectors. In previous surveys in the region, only a small percentage of livestock owners were aware of RVFV and not of any other arboviruses [
30,
31]. It is possible that several arboviruses affect cattle in this region. Thus, the objective of our study was to improve the knowledge of IEP circulation of RVFV and the subclinical occurrence of BTV and EHDV in cattle in Baringo County, as well as to contribute data to a better understanding of the distribution of these viruses in diverse ecosystems in the Horn of Africa.
4. Discussion
To advance the knowledge of RVFV circulation during IEPs, the epidemiology of BTV and EHDV, and the associated risk factors, this serosurvey was carried out in Baringo County, Kenya, an RVF high-risk region. We found evidence of past RVFV exposure in cattle based on the presence of IgG antibodies against RVFV. While no IgM antibodies were detected, the presence of IgG antibodies in young cattle (6–12 months) and the fact that all the tested animals were born after 2007 (when the last recorded outbreak occurred in Baringo) indicate the cryptic circulation of RVFV. Furthermore, we report very high seroprevalence rates of BTV and EHDV. From the results of risk factor analysis, the age of the animal was one of the factors associated with exposure to all three viruses. Surveillance of RVF during IEPs helps in understanding its circulation, as it is generally thought to contribute to the maintenance of the virus in vertebrates and its replenishment in mosquito populations [
9]. It is also important to define epidemiological zones of other relatively unknown viruses, which in turn informs control efforts against them.
The seroprevalence of RVFV reported here (15.5%) was lower than that previously reported in Tanzania (29.2%) [
49] and South Africa (42.9%) [
50] during IEP periods. It was, however, higher than the reported seroprevalence in Ijara, Kenya (13.1%) [
51] and Kilombero, Tanzania (5.5%) [
52]. In the 2006/2007 major outbreak in eastern Africa, both Kenya and Tanzania were affected, with almost similar human and livestock cases reported [
53]. In a longitudinal study in Ijara County, higher seroprevalences were reported when migrating herds passed through thick forests compared to those grazing near homesteads. Yet, in our study, the distance of the household/herd from the nearby community wildlife conservancies was not associated with RVFV seropositivity. The presence of IgG antibodies against RVFV indicates previous exposure; however, the oldest animal sampled in this study was 15 years old, implying that all the animals sampled in our study were born after the 2006/2007 major RVF outbreak in East Africa. Thus, it can be assumed that all seropositive animals detected in Baringo County had acquired the virus thereafter. Recent (2021) smaller outbreaks of RVF in Kenya occurred in other counties such as Isiolo, Mandera, Garissa, and Marsabit [
47,
54], and these counties do not border Baringo. This limits the possibility that the recorded virus circulation was derived from RVFV-infected animal incursions. Given that vaccination against RVFV in this region was last conducted during the 2007 outbreak and maternally derived antibodies decline after 6 months, seropositivity in animals between 6 to 12 months also suggests recent exposure to the virus. Moreover, modeling studies have shown a decline in immunity in cattle within 5 years following an outbreak [
55].
The seroprevalence of BTV (91.5%) in this study is similar to other studies in Kenya and beyond, while the EHDV seroprevalence (91%) we recorded is higher than those reported in previous studies in Kenya and elsewhere. For instance, in western Kenya, 94% and 64% of indigenous calves had seroconverted to BTV and EHDV by 51 weeks of age, respectively [
21]. In Machakos County, Kenya, 50% of the sampled cattle were seropositive for BTV while in Isiolo County, a seropositivity more similar to our study was reported (86%) [
56]. These regional differences might arise due to differences in the ecology, which affects the breeding and activity of the vectors. Elsewhere, varying seroprevalence rates against BTV in cattle were reported, with 99.5% in Mayotte [
57], 88.8% in Mali [
17], 19.4% in Sudan [
58], and 62% in Zimbabwe [
59]. In the Campania region of Italy, a prevalence of 45.2% was observed in cattle, while in Pakistan, 66% of the tested cattle were seropositive for BTV [
60,
61]. Even though in most of the epidemiological studies involving BTV/EHDV, the tested animals do not exhibit any clinical illness, they reflect the importance of cattle in the maintenance of the viruses.
We noted a high seroprevalence of BTV, similar to most of the previous studies, emphasizing a very high infection rate in cattle which is probably driven by the presence and activity of
Culicoides spp. and subclinically infected cattle in the tested herds. Baringo County is a semi-arid region, hence most of the cattle graze on the shores of Lakes Baringo and Bogoria, around irrigation plots and in swamps in the nearby wildlife conservancies. These areas are perennially wet and muddy, conditions highly conducive for
Culicoides spp. [
62]. Although the high temperatures and drought-like conditions experienced in the county reduce vector survival, they create shallower and warmer water bodies, exposing mud which favors the breeding of midges. Likewise, these conditions also push cattle to aggregate at watering points for most of the day, exposing them to biting midges [
63].
Previous studies have shown that the epidemiological features of BTV are likely to mirror those of EHDV as they share the same vector and ecological niche. However, exposure to EHDV is consistently lower than to BTV in most of the studies, underlining subtle differences in the epidemiology of BTV and EHDV. Insignificant differences have been reported in the replication rate, temperature required, and extrinsic incubation period of BTV, EHDV, and African horse sickness virus in
C. sonorensis [
64,
65]. We therefore hypothesize that the difference in seroprevalence is a result of specific virus strain–vector species relationships. It is possible that there are differences in vector competence for different orbiviruses in the same vector species and for the same orbivirus in different vector species [
64]. A recent entomological survey in Baringo County found
C. imicola to be the most prevalent species [
66], hence relevant information could be generated by serotyping the local BTV and EHDV strains and then using them for transmission experiments in
C. imicola.
The seroprevalence of EHDV in our survey was lower than that in a cross-sectional study of EHDV in cattle in Mayotte (96.9%), possibly due to the previous outbreaks of both BTV and EHDV in Mayotte [
57]. In a case-control study in dairy cattle in Israel, the seroprevalence of EHDV was reported to be 72% in clinically affected cattle while there was also a higher prevalence in subclinical animals in the herd (57%) [
67]. In Zimbabwe, a seroprevalence of 57% was reported in cattle towards the end of the rainy season. This seroprevalence was higher than that recorded during the dry season [
59]. In summary, several factors including wetland cover, soil type, temperature, livestock management system, vector control, and wind patterns influence the exposure of animals in a given locality, leading to the reported differences in seropositivity to BTV and EHDV [
58,
68].
There was a significant correlation between BTV and EHDV seropositivity, which was absent between RVFV and each of the two viruses. This dual exposure/co-infection has been described in previous studies and is plausible given that BTV and EHDV usually share the same
Culicoides vector and transmission ecology [
21,
57,
62]. High seropositivity to these two viruses in cattle in the absence of clinical cases implies an endemic situation in Baringo County and most likely the rest of Kenya. It also signifies natural infection as the Directorate of Veterinary Services in the county confirmed that no vaccination had been carried out against the two viruses (personal communication).
In addition to the loss of body condition, abortions, and decreased milk production associated with a subclinical disease state, other serotypes of EHDV (EHDV serotypes 2, 6, and 7) [
24] and BTV (BTV serotype 8) [
69] are also known to cause overt clinical signs in cattle and possibly death. Control efforts should therefore be implemented in endemic zones where there is high
Culicoides spp. activity, such as in Baringo County [
66]. These measures generally include vaccination with the dominant BTV strains in the area and keeping livestock off low-lying wet grazing lands. In Kenya, a BLUEVAX™ freeze-dried, live attenuated bluetongue vaccine prepared from the seven common strains is available on the market for yearly sheep vaccination [
70]. However, widespread use of the vaccine is hampered by the cost and logistical reasons. On the other hand, vaccination for EHDV is carried out only in Japan and the USA, where the disease causes significant losses in white-tailed deer [
24].
Several key variables were investigated that are likely to affect/associate with the seropositivity of the three viruses. The age of the animal was significantly associated with seropositivity to all three viruses, indicating that adult animals were more seropositive compared to young animals/calves. For RVFV, age-dependent exposure has been reported previously and is compounded by the longevity of IgG antibodies in cattle. Therefore, the older the animal, the more chance it has to be infected by the virus [
49,
71,
72]. The same trend is also evident for BTV and EHDV, where older cattle are likely to be more frequently exposed to
Culicoides bites compared to the young [
60,
61,
73]. We found female animals were significantly more infected with RVFV compared to males. This could be because most of the females stay in the herd longer for reproductive purposes and herd growth. On the other hand, bulls and oxen are usually sold off quickly for slaughter or social ceremonies [
71]. With respect to the animal owner, an older age (>60) was associated with seropositivity compared to those between 30–39 years. This is a linkage that has not been investigated in previous studies and could possibly be attributed to better knowledge and livestock management practices in younger compared to older cattle owners.
We also report low ICC values for the three viruses, which can help in the design of future studies. With respect to RVFV, the ICC was lower than that reported in Tana River County, Kenya [
35] and in South Africa [
50]. This shows that there was less between-herd variation in seroprevalence to RVFV in Baringo County compared to Tana River County and South Africa. A low ICC is reflective of conditions that are more conducive to IEP circulation and widespread exposure. A high ICC therefore implies that more herds should be used during epidemiological studies with fewer animals sampled per herd. Overall, one of the dangers of clustering is a misrepresentation of the prevailing risk of RVF or any other disease [
35].
A limitation of our study is that some recruited households had fewer cattle than previously indicated, leading to the recruitment of additional households on sampling day and more than 10 animals being sampled per herd. Additionally, viral neutralization tests were not carried out on the ELISA-positive samples to ascertain the presence of neutralizing antibodies. However, previous studies have shown that the difference between the ELISA test and neutralization tests is minimal, with 95–100% of ELISA-positive samples confirmed to have neutralizing antibodies [
49]. The IDvet ELISA kit that we used has been validated and its sensitivity and specificity were reported to be 91–100% and 100%, respectively [
39]. While there is a possibility of cross-reaction between the closely related BTV and EHDV, the IDvet ELISA kit that we used showed 100% specificity for EHDV in the presence of BTV in test serum samples [
41]. Despite these limitations, we are confident that the seropositivity reported in this study is a true reflection of the exposure to RVFV, BTV, and EHDV in cattle in Baringo County, Kenya.