Climate,environment and disease

1 Background

Rift Valley fever (RVF) is a viral zoonosis present in ruminants, camels, and humans, and is found throughout much of Africa. Spread via mosquito vectors, as well as fluids and tissues from infected animals, RVF is an arbovirus associated with illnesses such as headache, myalgia, and liver abnormalities, and has in a very small percentage of cases (1–3%) resulted in hemorrhagic fever. The virus was likely first identified in 1931 by Daubney et al. (1931) in Kenya, but has since spread throughout much of the African continent. In 2000, outbreaks were recorded in both Yemen and Saudi Arabia, marking the first RVF outbreak recorded outside of Africa (Flick and Bouloy, 2005). As the effects of global change begin to alter climatic and ecological processes throughout the world, the emergence of diseases and their vectors are certain to shift as well. This progress report addresses this topic through a review of research throughout the endemic area, and includes consideration of climate change as well as new work published in recent years. In particular, it highlights the intersections of climate-driven environmental change with RVF and suggests new directions for research in the future.

2 Disease ecology and epidemiology of Rift Valley fever

Transmitted by mosquitoes of the genera Aedes and Culex (Métras et al., 2011), RVF is a viral disease that can infect both animals and humans. RVF has been documented in numerous animals, including sheep, goat, cattle, and water buffalo (Mandell and Flick, 2011), often resulting in high rates of spontaneous abortion, severe disease, and death (Pfeffer and Dobler, 2010). As a result of large-scale infection in livestock, RVF can devastate local economies and critically threaten public health. Human infections are typically associated with farming and livestock rearing, therefore primarily affecting rural and peri-urban dwellers (WHO, 2010).

Before the late 20th century, RVF was generally found in parts of eastern and southern Africa (CDC, 2007). However, since the 1977 documentation of a large outbreak affecting 18,000 people in Egypt, epidemics have been increasingly reported throughout Africa. As of September 2000, RVF has also been reported outside of Africa (Figure 1). In September 2000, the first RVF epidemic in Saudi Arabia and Yemen was reported, resulting in 886 cases of human disease (Madani et al., 2003). Although speculation remains as to the possible causes of this geographic leap in transmission, researchers hypothesize two causes: the migration of infected camels, or the encroachment of mosquitoes into new areas (Gubler, 2002; Pfeffer and Dobler, 2010). The hypothesis of livestock movement is supported by the genetic similarity of virus samples collected in Saudi Arabia to a strain isolated in Kenya during the 1997–1998 outbreak (Chevalier et al., 2010).

OPEN IN VIEWER

Figure 1.

Geographic distribution of RVF outbreaks. Dark red shading indicates areas with reported outbreaks and light red shading indicates areas at risk for RVF based on reported evidence. (See colour version of this figure online).

This geographic leap of RVF is indeed a cause for alarm as it illustrates the potential of the disease to spread rapidly to large populations of humans and domestic animals (Clements et al., 2007; Mandell and Flick, 2011). The ease through which the virus can be transmitted to varied hosts through diverse mosquito species has led the United States to recognize the virus as a threat and the European Union to undertake an RVF risk analysis (Chevalier et al., 2010; Mandell and Flick, 2011). In the United States, RVF virus is categorized as an overlap Select Agent as it is believed to ‘pose a severe threat to animal health, animal products, and public health’ (NSAR, 2011).

The virus is transmitted to humans by mosquitoes as well as via direct or indirect contact with fluids from an infected animal. Transmission via fluids includes direct contact between a human wound and animal blood or organs, or through inhalation of aerosol particles. Although primarily a livestock disease affecting domestic animals, RVF outbreaks result in a wide array of symptoms often leading to severe disease or complications in 2% of cases in humans (Mandell and Flick, 2011). For example, an estimated 20,000 human cases were documented in the 1951 outbreak in South Africa (Woods et al., 2002). The disease may serve as an occupational hazard for herders, butchers, veterinarians, and farm workers who are in regular contact with animals (WHO, 2010).

In Africa, 23 mosquito species are responsible for not only maintaining the epizootic cycle but also initiating human epidemics (Flick and Bouloy, 2005). Although disease-transmitting mosquito species differ from region to region (i.e. eastern, western, or southern Africa) and exhibit different biting behavior, many RVF-spreading species share the same type of breeding sites and feed on the same hosts, such as cattle and sheep (Fontenille et al., 1998; LaBeaud et al., 2010). Floodwater mosquito species such as Aedes have been shown to be capable of both vertical and horizontal transmission of the virus, the former maintaining the virus in nature during inter-epidemic time periods (Gubler, 2002; Wilson, 1994), and the latter transmitting the disease between animals. For this reason, outbreaks are found to be associated with periods of heavy rainfall and flooding (Gubler, 2002; Pfeffer and Dobler, 2010), as well as the construction of dams (Flick and Bouloy, 2005). It must be noted that once conditions change and standing water is available for more than two to three weeks, a transition of vectors can occur, allowing for the proliferation of other mosquitoes, such as Culex, which further amplify the outbreak (Pepin et al., 2010). However, Pepin et al. assert, this transmission cycle is more indicative of outbreaks occurring in areas of sub-Saharan, eastern, and southern Africa while RVF outbreaks in western Africa and the Arabian peninsula may be associated with other species of mosquitoes with a preference for permanent bodies of water, such as dams.

RVF outbreaks are detected in animal populations before cases are reported among humans, making it important to implement disease prevention and control programs quickly. Table 1 summarizes the recommendations of the World Health Organization (WHO) for the control and prevention of RVF in animals and humans. However, it is difficult to implement prevention programs quickly to mitigate the spread of disease due to the lack of a comprehensive disease surveillance system or an early warning system. Currently, the Smithburn vaccine is the only approved effective livestock vaccine for use during outbreaks (Boshra et al., 2011). However, due to the delay in onset of immunity, there is much speculation that the vaccine has limited effectiveness once an outbreak has already started, making it necessary to use a multi-pronged strategy as part of the larger disease control program (Breiman et al., 2010). Although select vaccines under development are showing promise, challenges that need to be overcome include cost-efficiency, providing protection in a single dose, eliminating side effects, and ease of production and administration (Boshra et al., 2011; LaBeaud et al., 2010). The absence of a licenced, commercially available human vaccine makes it very difficult to prevent disease transmission in humans once an outbreak is underway, making it necessary to implement integrated disease control programs, including community-based education campaigns, to prevent the spread of RVF.

OPEN IN VIEWER

Table 1.

Summary of recommendations from the World Health Organization (2010) for the control and prevention of RVF in animals and humans.

Recommendations	Animals	Humans
Prevention (before an outbreak)	Sustained vaccination program	Implement public health education programs to reduce risk of disease, including the use of personal protective measures (e.g. gloves, protective clothing) for high-risk populations (i.e. butchers, animal husbandry, etc.) to minimize contact with animal blood or body fluids; use of bed nets and insect repellents to prevent mosquito bites
Control (during an outbreak)	Restrict or ban movement of animals to prevent virus from spreading to other locations	Continue to educate and sensitize community members to risk factors associated with RVF; in addition to the above preventative methods, encourage the use of protective measures against mosquitoes (e.g. larviciding, bed nets, restricted outdoor movement during peak biting times), and safe animal handling and cooking practices

In terms of RVF detection, key indicators that can signal a potential outbreak include a spike in abortions among domestic animals, heavy rainfall, and an increased number of mosquitoes. However, in order to improve RVF surveillance – or, better yet, to devise an early warning system that may help prevent outbreaks – a more thorough understanding of the relationship between RVF, climate, and ecology is critical.

1 Regional climate variability and outbreaks

The World Health Organization (WHO, 2010) notes the distinct connections between RVF outbreaks and regional climate variability, particularly those of rainfall patterns. This connection has been highlighted in a number of studies in Africa (e.g. Anyamba et al., 2002; Davies et al., 1985; Linthicum et al., 1987). These rainfall events are often embedded within regional climate circulations, including the Intertropical Convergence Zone (ITCZ) in the equatorial region. The ITCZ is characterized by a band of thunderstorms and mesoscale convective systems that track north and south of the equator throughout the year as a result of solar heating, and generally produces one or two annual rainy seasons for the areas below.

RVF outbreaks are also influenced by the El Niño/Southern Oscillation (ENSO) teleconnection (Anyamba et al., 2002, 2010) that leads to varying climate impacts across much of Africa on roughly a three- to seven-year timescale. Though geographically variable, the El Niño phase of ENSO produces increased rainfall in East Africa (Nicholson, 1996), while wetter years in the southern portion of the continent appear to follow La Niña events, tied to changing sea surface temperatures (Nicholson and Selato, 2000). Subsequently, this pattern has broadly been linked to RVF outbreaks in equatorial East Africa during the El Niño phase and in southern Africa during the La Niña phase, in concert with the Indian Ocean Dipole (Anyamba et al., 2001, 2010). The inclusion of large-scale climate analyses has also enabled the identification of regional hot-spots, such as portions of Senegal and Mauritania within West Africa as well as Madagascar (Anyamba et al., 2010; Caminade et al., 2010).

The connection of these heavy rains with epidemic outbreaks results from the ecologies of the many potential RVF mosquito vectors (Meegan and Bailey, 1988). Floodwater Aedes mosquitoes, often the dominant vectors, are well adapted to this periodic rainfall. Following rains, breeding is increased due to ample habitat. Because RVF can be passed transovarially from infected female mosquitoes to their eggs (Davies et al., 1985; Linthicum et al., 1983, 1985), mosquitoes can hatch already infected with the virus, potentially transmitting RVF through bloodmeals. As Martin et al. (2008) note, during dry periods the eggs can remain viable in the mud for several years until heavy rains flood the soil and infected mosquitoes emerge, triggering epidemics. In some instances, other vectors such as Anopheles and Culex mosquitoes are responsible for or add to RVF transmission (Martin et al., 2008; Meegan and Bailey, 1988).

Yet, given the large geographic coverage of RVF outbreaks in Africa (Figure 1), it is important to note the limited spatial coverage of ecological RVF case studies. In their review of RVF modeling studies, Métras et al. (2011) note that the majority of RVF risk studies were conducted either in Kenya or West Africa (e.g. Senegal or Mauritania), while most regional climate studies of RVF were based upon ecological relationships observed in Kenya. Therefore, when extrapolated across Africa, it is not surprising that the best spatial predictions were for East Africa, with less successful predictive performance for Sudan, Madagascar, and southern Africa (Métras et al., 2011). Despite previous research (e.g. McIntosh et al., 1980), the limited amount of recent published research on ecological relationships in southern Africa raises concerns given the endemic foci of RVF there and the different atmospheric patterns dominating the climate of the region. Additional investigation could increase understanding of how climate may affect RVF in locations farther from the equator.

2 Future climate change impacts on RVF outbreaks

While it is expected that temperatures throughout Africa will increase under climate change, Intergovernmental Panel on Climate Change (IPCC) (2007) projections of rainfall are less certain and vary by region, although it is likely that both wet and dry years will become more extreme. Thus the cyclical production of floodwater Aedes eggs infected with RVF is likely to continue in the future. However, a possible change in the length of time between these wet and dry years may influence the periodicity of epidemics (Martin et al., 2008).

While the linkages of RVF and precipitation are frequently addressed, less attention has been paid to the role of changing temperature, though it has been posited as a possible variable to consider (Martin et al., 2008). Temperature has been noted as an influence on the incubation period and transmission of the RVF virus. Turell et al. (1985) demonstrated that, at higher temperatures, greater numbers of Culex pipiens were infected, and the length of time that it took for the mosquito to become infectious was shortened in both Aedes taeniorhynchus and Culex pipiens. While laboratory-based experiments cannot be expected to completely mimic ecological processes, additional work by Turell (1989) further established the connection between temperature and RVF transmission in Aedes fowleri using a study design more attuned to natural temperature patterns than previous studies. However, there is room for further research in this regard, as little work has been done on the future of Rift Valley fever under changing temperature, particularly compared to other vector-borne diseases such as dengue and malaria. Often, RVF is only considered within broader reviews of climate change and health (e.g. Epstein, 1999; Githeko et al., 2000; Patz et al., 2005), or through climate change and RVF-specific studies focused predominantly on precipitation (Martin et al., 2008, Tourre et al., 2009). We suggest that given the localized focus of climate-RVF case studies to date, it is crucial to explore these relationships in new locations. Otherwise, predictive studies for other places using climate change projections may run the risk of error or bias because the underlying disease-climate relationships may not be generalizable.

3 Research orientation for adaptation and mitigation

The IPCC includes RVF as one of numerous infectious diseases to be influenced by fluctuations in the climate system. In particular, a potential rise in El Niño events is expected to exacerbate the spread of RVF through associated rainfall increases in some locations (Parry et al., 2007). Climate change will produce spatially and temporally uneven precipitation and temperature events, both in Africa and across the globe, not to mention the correspondingly uneven impacts of these trends on diverse populations (Parry et al., 2007). While adaptation and mitigation are central to addressing climate change and its impacts, they can be problematic because certain climate events may harm society in some ways but benefit it in others. For example, increases in precipitation may cause more RVF outbreaks but, at the same time, can benefit human health through increased water availability and improved food supply.

The language of mitigation is often used in relation to the control and prevention of diseases such as RVF. It is therefore useful to make the distinction between RVF mitigation and climate change mitigation: the former involves spatially, geographically, and culturally nuanced strategies for strengthening vector control, host resilience, and overall awareness about the practices that cause animal-to-human infection (the leading mode of transmission); the latter involves reducing greenhouse gas emissions to curb escalating climate variability, extremes, and their repercussions for human and ecological systems. This distinction is crucial, because RVF is not just another dependent variable of climate – and climate change – to which vulnerable populations must adapt. Further insight is needed into both the climatic and non-climatic components that shape the disease, its vectors, and the vulnerability and resilience of its hosts.

1 Remote sensing

Remote sensing approaches can provide localized information on environmental factors that may be associated with disease-vector habitats and human transmission risks (Beck et al., 2000). Specifically, satellite monitoring has made it possible to monitor vegetation conditions associated with epidemiological outbreaks (Anyamba et al., 2001). Spatiotemporal changes in environmental patterns can be tracked via remote sensing to identify geographic areas with specific conditions related to disease outbreaks (Anyamba et al., 2009). For RVF research, moderate resolution satellites such as Landsat and the Système Pour l’Observation de la Terre (SPOT), have been used in conjunction with airborne synthetic aperture radar in East Africa, not only for monitoring the disease but also to identify potential RVF vector breeding sites and their spatial distribution (Anyamba et al., 2001).

The Normalized Difference Vegetation Index (NDVI) has been used to measure seasonal trends in photosynthetic vegetation activity (Tucker, 1979), which are in part a response to climate. This index, in combination with other climate variables, has been used to map areas in Kenya where RVF has occurred (Anyamba et al., 2002, 2006; Boast, 1990; Hightower et al., 2012; Linthicum et al., 1987, 1990; Logan et al., 1991). Pope et al. (1992) used Advanced Very High Resolution Radiometer (AVHRR) analyses to predict RVF outbreaks in Kenya based on NDVI values, focusing on conditions suitable for the earliest stages of a RVF epizootic outbreak.

Evidence accumulated from various studies related to diseases in Africa has shown that there is a close relationship between green-leaf biomass development and breeding and expansion patterns of some insect pests and disease vectors (Linthicum et al., 1987, 1990; Tucker et al., 1985). Linthicum et al. (1999) demonstrated that above-average seasonal rainfall in East Africa stimulates a seasonal increase in RVF and other vector-borne epidemics. The same relationship between environmental variability and RVF outbreaks was observed in the Ferro region of Senegal, West Africa, by using SPOT imagery with 10 m pixels (Vignolles et al., 2009). Similarly, Hightower et al. (2012) have linked increased greenness and prior rainfall events to RVF outbreaks and associated local soil types and land cover to increased risk of disease. This close relationship between rainfall, vegetation growth, and mosquito life-cycle dynamics can be quantified using satellite time series measurements, and may help map and predict areas with high risks for RVF activity. Furthermore, as Kalluri et al. (2007) illustrate, remote sensing techniques are already being utilized to study other vector-borne diseases such as African Trypanosomiasis and Lyme disease. Of the other prevalent diseases in Africa, African Trypanosomiasis is most relevant due to its similar impact on livestock populations in many of the same regions as RVF, spanning from central and western Africa, for Trypanosoma brucei rhodesiense, to eastern and southern Africa, for Trypanosoma brucei gambiense (Fèvre et al., 2006). Several papers provide good references regarding the applications of remote sensing to model vector populations (Rogers, 2000; Rogers et al., 1996).

A major concern for monitoring any epidemiological outbreaks with remote sensing data is the incongruence between timelines of data acquisition and vector life cycles (Pope et al., 1992). In addition to this is the inability of remote sensing to capture conditions related to the other form of RVF transmission through human exposure to contaminated animal fluids. While remote sensing may be useful for monitoring dynamics conducive to mosquito vector development at broad landscape scales, it accounts for neither the dominant mode of animal-human transmission nor the myriad other smaller-scale processes that enable outbreaks. For this reason, it is necessary to utilize additional monitoring and management tools for optimal identification of RVF outbreaks in both animal and human populations.

2 Other approaches

There are a number of additional approaches that can strengthen environmental research on RVF and its control in the future. Generally, these options are more geographically diverse, finer-scaled, and locally attuned to the interconnected environmental and social factors that enable RVF outbreaks. In order to understand local drivers of zoonotic diseases, Bender et al. (2006) recommend examining the social and economic idiosyncrasies of outbreak sites. In Yemen, for example, the first recorded outbreak of RVF in 2000 coincided with the Islamic Eid al-Kabeer festival, a period of high densities of animal hosts (cattle, sheep, and goats) imported primarily from the Horn of Africa (Abdo-Salem et al., 2011a). Added to the importance of socio-economic factors is a need to better understand the interrelations of pathology, ecology, and sociology that shape zoonosis management (Breiman et al., 2008). Kimani et al. (2009) also prescribe committed, in situ collaboration between public health, education, agriculture, and economic sectors to ensure consistency and continuity in surveillance, prevention, and control efforts.

Attention to the climate-RVF link tends to focus narrowly on the positive correlations between ENSO-related rainfall events and outbreaks (Anyamba et al., 2001; Martin et al., 2008). Based on the underlying disease ecology and epidemiology, more nuanced analyses are thus needed to trace the multitude of factors that underpin these correlations. In particular, it is not clear what micro-scale and meso-level processes converge to form this seemingly stable broader signal between ENSO and RVF. Moreover, how might variations in each alter the overall relationship between precipitation and epidemiology at a particular site? This need to understand finer-scale processes extends to RVF outbreak and temperature interactions, as well as to the expansion of RVF research to other epidemic and endemic regions, particularly the Middle East and southern Africa, as mentioned previously.

Livestock migration and trade could be a key entry point for understanding how the disease spreads between regions (Abdo-Salem et al., 2011b). The influence of climate on animal-human RVF transmission rates has yet to be explored. Given that the majority of RVF cases are caused by animal-human transmission, improved understanding of the relevant socio-economic factors coupled with knowledge of climate-RVF relations would inform locally adapted strategies for RVF control.