As natural hosts for avian influenza virus (AIV), wild birds, particularly aquatic birds, are the primary reservoir for transmission of AIV to domestic poultry. Therefore, understanding the dissemination and maintenance of AIV in wild birds is important for understanding the factors that contribute to transmission of AIV from wild birds to poultry. However, relatively little is known about the ecology of the virus in wild birds, and the depth of data are inconsistent worldwide. Also, the biology of the virus itself is very important, because AIV is highly adaptable to different hosts and likely to the environment as well. Some insight may be gained from the Asian H5N1 highly pathogenic AIV lineage, because surveillance for this virus has increased considerably in wild birds worldwide since 2005. Also, numerous species that have not previously been represented in AIV testing have been included in surveillance for the Asian H5N1 highly pathogenic AIV, allowing for a more complete understanding of the distribution of AIV in wild birds.
DISEASE ECOLOGY, HOST BIOLOGY, AND AVIAN INFLUENZA VIRUS DISSEMINATION AND MAINTENENACE
Stallknecht and Brown (2008) outline 3 factors in the transmission and maintenance of avian influenza virus (AIV) in its natural reservoirs: 1) virus shed at sufficient titers for an adequate duration of time, 2) the environmental stability of the virus, and 3) the titer required for productive infection of the next host (infective dose). These factors, which vary among host species and habitats, fit together to form the ecology of AIV.
The ecology of AIV is inherently tied to the biology of the natural host and environment. For example, mallard ducks (Anas platyrhynchos) and many of the other duck species with relatively high prevalence rates of AIV infection are dabbling ducks (genus Anas) who live in freshwater habitats in contrast to shorebirds and gulls, which are saltwater birds. Characteristics like these may appear to be incidental for the virus, but work done by Stallknecht et al. (1990) demonstrates that some AIV isolates are more sensitive to salinity than others. Because transmission probably occurs by the indirect fecal-oral route, the virus must be stable in the environment. Other host characteristics that may affect AIV ecology are feeding habitats, migration, mating behavior, and maintenance of a minimum population size of susceptible individuals.
It should also be recognized that humans may both directly and indirectly affect the ecology of AIV. For example, the role of wildlife trade (both legal and illegal) is unknown and may provide an artificial interface among numerous domestic avian and mammalian species. In some situations, these animals are released back into the wild (e.g., merit release), providing an additional interface with wild birds (Karesh et al., 2005). Other anthropogenic effects on domestic and wild bird movement may have more long-term significance (for example, climate change and urban development), which can confound disease epidemiology by affecting migration patterns and stop-over-staging places, population densities, and the comingling of different species (reviewed in Reed et al., 2003).
Finally, there is the question of how disease caused by AIV in reservoir species may affect virus spread. Normally, neither low pathogenic AIV (LPAIV) nor viruses that are classified as high pathogenic AIV (HPAIV) for gallinaceous birds cause disease in ducks (Swayne and Halvorson, 2003). Pathogenesis data on highly pathogenic avian influenza in gulls and shore-birds are largely absent except for some studies with several strains of the Asian H5N1 HPAIV. Perkins and Swayne (2002) looked at a 1997 strain, which showed disease in gulls to be mild; however, severe disease and mortality in laughing gulls were reported with experimental infection with strains from 2001 and 2006 (Brown et al., 2006). Historically, only 2 AIV lineages (both which have been classified as HPAIV for gallinaceous birds) have been associated with disease in natural host species: 1) A/Tern/SouthAfrica/61 (H5N3; Becker, 1966) and 2) some strains of the Asian H5N1 HPAIV lineage that emerged around 2003. Therefore, AIV-induced disease does not exert a major influence on the virus ecology in natural host species.
The most critical factor for AIV dissemination and maintenance is probably viral shedding (concentration, duration, and route). Viral titers and duration of shedding can be affected by virus strain, bird species, bird age, and overall health status. The route of viral excretion may also be a factor. For example, in contrast to LPAIV and most HPAIV, the Asian H5N1 HPAIV are shed at much greater titers and for longer durations from the trachea and upper respiratory tract than by the cloacal route (Tumpey et al., 2002; Kishida et al., 2005; Pantin-Jackwood et al., 2007). This deviation in route of shedding may affect the risk for viral contamination of the aquatic environment by infected birds and therefore may influence the potential and efficiency of virus transmission within the population or to domestic poultry (Stallknecht and Brown, 2008).
RESERVOIR SPECIES AND GEOGRAPHIC DISTRIBUTION
Based primarily on greater AIV prevalence rates, several species of dabbling ducks [mallards (A. platyrhynchos), Northern pintails (Anas acuta), blue-winged teal (Anas discors), and green-winged teal (Anas crecca)] and species in the order Charadriiformes [ruddy turnstones (Arenaria interpres)] and a variety of gull species (family Laridae) have been identified as reservoirs for AIV. However, relatively little is understood about the global ecology of AIV and the full range of host species that contribute to the maintenance of virus. Overall, AIV has been isolated from more than 100 species of 12 orders of birds (reviewed in Stallknecht and Brown, 2008), many of which are considered to have been infected by contact with infected poultry and are currently not considered important in the ecology of AIV. However, additional species, which have yet to be identified, may play a critical role in the ecology of AIV.
Furthermore, sampling bias may exist in the current data regarding the prevalence of AIV in different wild birds and which species are the natural hosts, because certain geographical regions, species, and seasons have been favored due to accessibility and other logistical concerns. For example, a species may be preferentially sampled because they can be caught or are in high concentrations in accessible areas. Additionally, not all wild bird monitoring is done with the same virus detection methods (e.g., reverse transcription PCR versus virus isolation versus serology).
Although AIV is found worldwide, the depth of data on AIV distribution varies considerably. The most comprehensive and long-term data are from North America and Europe, with somewhat less data from Asia and Oceania and the least data from Africa and South America. Importantly, virus subtype distribution and virus prevalence vary by bird species, age, time of year, and geographical location.
Asian H5N1 HPAIV and Wild Birds
The Asian H5N1 HPAIV is unprecedented among modern AIV outbreaks in the breadth of its geographical distribution. Between 2003 and early 2008, sixty-one countries have reported detection of the virus in domestic poultry or wilds birds, 13 of which reported the virus in wild birds only (OIE, 2008). The emergence of the Asian H5N1 HPAIV has led to increased monitoring of wild birds worldwide for AIV, and current data are being generated for many species and regions worldwide that had not been previously monitored or sampled. The Asian H5N1 HPAIV lineage is also very diverse. It has undergone numerous reassortment events (which is indirect evidence of transmission among wild birds) and continues to accumulate point mutations within genes, analysis of which can help establish routes of geographical spread relative to other sublineages of the H5N1 viruses.
Although this AIV lineage may provide some valuable information on the dissemination and spread of AIV by wild birds, one must be cautious in extrapolating the data because of the unusual biological characteristics of this virus lineage including: 1) the ability of some strains to cause disease and death in some species of waterfowl (Sturm-Ramirez et al., 2004), 2) tissue tropism for the respiratory tract rather than the intestinal tract (Tumpey et al., 2002; Kishida et al., 2005; Pantin-Jackwood et al., 2007), 3) a wide distribution among domestic poultry, and 4) unusual transmissibility to numerous mammalian species not normally associated with AIV (e.g., felids and stone martens; Kuiken et al., 2004; Starick et al., 2008). Another consideration is that testing procedures are not standardized and often the approach is to look specifically for the Asian H5N1 HPAIV versus testing for any and all AIV. Consequently, the data for AIV in general may be incomplete.
Infections with the Asian H5N1 HPAIV were first documented in wild birds in Hong Kong around 2002, where the infections probably initially occurred as spill-over from infected poultry (Ellis et al., 2004). Sometime around late 2005, the virus began to spread out of Southeast Asia into Russia, Europe, and Africa. Although the mechanism of virus dissemination is difficult to definitively establish, there is evidence that wild birds were involved as well as movement of poultry and domestic birds (reviewed in Sims and Brown, 2008).
The Asian H5N1 has been isolated from wild birds in Europe in 2006, 2007, and 2008, and importantly, most of the detections in wild birds have been from dead birds (Starick et al., 2008) and were discovered through surveillance efforts (Brown et al., 2007). The viruses isolated from poultry and wild birds in Germany (Starick et al., 2008), France (Gall-Recule et al., 2008), the Czech Republic (Nagy et al., 2007), Nigeria (Monne et al., 2008), and Russia (Lipatov et al., 2007) were related to Qinghai Lake viruses (Clade 2.2), suggesting that the virus spread from Asia into Europe and Africa by a single route or from a the area around the lake. All these outbreaks were short-lived, indicating that the virus was not maintained in bird populations. More concrete evidence of transmission of viruses from wild birds comes from studies in which numerous genes were analyzed, thus revealing that reassortment is occurring (Salzberg et al., 2007; Monne et al., 2008). Reassortment indicates that these viruses are not being transmitted as discrete lineages, either among domestic or wild birds.
In Asia, the H5N1 HPAIV has been isolated from a more diverse range of avian species than elsewhere (reviewed in Stallknecht and Brown, 2008), which is probably a reflection of the virus being endemic in the region and of a good interface between domestic and wild birds because of the nature of the agricultural systems. Most of the nonaquatic wild bird species that have been affected are scavengers and peridomestic birds (e.g., sparrows, crows, and pigeons) that live commensally with humans and thus around poultry facilities (reviewed in Stallknecht and Brown, 2008). Therefore, it is believed that many of these species are being infected through their habitat (i.e., spill-over from infected poultry) and do not represent species that are otherwise important for the ecology of AIV.
TRANSMISSION OF AIV FROM WILD BIRDS TO POULTRY
Transmission of LPAIV from wild birds to poultry has occurred numerous times, and wild birds are considered the sources of AIV for poultry. Even if a direct link is not established for a given incursion, virus genetic data and epidemiology are frequently the best indication that an LPAIV was introduced into poultry from wild birds unless a direct link can be found. Access to outside areas where wild birds congregate, for example, to eat the domestic animal feed [e.g., range-reared turkeys in Minnesota (Halvorson et al., 1985)], appears to be the most important factor. Similarly, contamination of poultry drinking water with wild waterfowl droppings may also be a means of exposure.
Importantly, wild birds almost exclusively carry LPAIV, and wild bird-origin LPAIV may be poorly infectious for chickens and turkeys. Minimum infectious doses reported for some wild bird isolates are on the order of 106 to 107 50% egg infectious doses/bird (Condobery and Slemons, 1992; Spackman et al., 2006; Swayne and Slemons, 2008), although there may be differences in 50% bird infectious dose between turkeys and chickens for a given isolate (Spackman et al., 2006). Therefore, the most critical factor for transmission of AIV between wild birds and poultry is probably the exposure interface where an adequate infectious dose for a given strain may be achieved.
Overall, because the quality and depth of data are inconsistent among poultry outbreaks, the true extent of wild bird transmission of H5N1 HPAIV to poultry is not known. Also, in the reported cases, it is not always possible to definitively determine whether wild birds were the direct source of infection for domestic poultry. Wild birds are often implicated based on either virus genetic data or lack of another known epidemiological link. Most reported cases involve facilities where the birds are reared either partially or totally outside, and most are backyard poultry (i.e., Food and Agriculture Organization sectors 3 and 4). Consistent with this, a primary risk factor for transmission of virus between wild and domestic birds is rice cropping with domestic ducks, in which the ducks are reared around rice paddies (Gilbert et al., 2008). When wild birds have contact with poultry, regardless of species, geographical region, or time of year, the possibility and risk of transmission increases substantially.
Currently, there are still numerous unknown factors in AIV ecology from both the virus and host perspectives. As more data becomes available about wild bird-origin LPAIV and the Asian H5N1 HPAIV, some of these questions can begin to be answered, which will contribute to improving existing prevention and control programs. However, from the practical perspective of transmission of AIV between wild birds and domestic poultry, many of the questions about host and virus ecology may be academic, because the interface between wild birds and poultry is probably the most critical factor and is a factor that can be controlled through biosecurity.
We gratefully thank Justin Brown (University of Georgia, Athens) and Colleen Thomas (Southeast Poultry Research Laboratory, USDA, Agricultural Research Service, Athens, GA) for reviewing this manuscript.