Facing the threat of equine influenza


email: debra.elton@aht.org.uk.


Despite the availability of vaccines, equine influenza virus (EIV) continues to pose a threat to the racing industry. The virus spreads rapidly in unprotected populations and large scale outbreaks, such as those in South Africa in 2003 and Australia in 2007, can cost billions of pounds. Like other influenza viruses, EIV undergoes antigenic variation, enabling it to evade antibodies generated against previous infection or vaccination. The UK has an active surveillance programme to monitor antigenic drift and participates in an international collaboration with other countries in Europe, Japan and the USA to select suitable vaccine strains. Selection is primarily based upon characterisation of the viral haemagglutinin (HA), the surface protein that induces a protective antibody response; this protein is an important component of commercial vaccines. In recent years vaccine technology has improved and diagnostic methods have become increasingly sensitive, both play a crucial part in facilitating the international movement of horses. Mathematical modelling techniques have been applied to study the risk factors involved in outbreaks and provide valuable information about the impact of vaccination. Other factors, such as pathogenicity, are poorly understood for EIV yet may play an important role in the spread of a particular virus. They may also affect the ability of the virus to cross the species barrier, as seen with the transfer to dogs in the USA. Severity of infection is likely to be influenced by more than one gene, but differences in the NS1 protein are believed to influence the cytokine response in the horse and have been manipulated to produce potential vaccine strains.


Equine influenza is a major cause of respiratory disease in the horse and, despite the availability of vaccines, continues to cause problems around the world for the Thoroughbred industry. In the UK, vaccination has been mandatory for competition horses since 1981, following the serious large scale outbreak in 1979. While some other horses are vaccinated, the majority of the equine population in the UK remains unvaccinated. In unvaccinated horses, clinical signs typically include a harsh dry cough, laboured breathing, loss of appetite, depression and lethargy. Horses are also highly susceptible to secondary bacterial infections, which may result in severe pneumonia. In vaccinated horses or those that have been previously infected, clinical signs are usually mild or unapparent. However, there have been instances of vaccine breakdown in the past, where the vaccine strains used have failed to protect against the circulating virus. It is therefore important that equine influenza vaccines are updated regularly. Much effort in recent years has been put into studying antigenic variation, the development of new vaccination strategies and, more recently, the factors that may be important for the development of disease. In the last decade, equine influenza virus has crossed the species barrier in the USA and become established in dogs. There have been sporadic cases in the UK and Australia, but fortunately the virus has not yet become established in other canine populations. This review will cover the evolution and pathogenicity of equine influenza viruses and importance of surveillance, diagnostics, current vaccines, the use of mathematical modelling for studying epidemics and the cross species transmission into dogs.

Antigenic variation

Influenza A viruses are essentially viruses of wild aquatic birds (Table 1). As they have evolved and changed over time they have gained the ability to cross the species barrier and become established in new hosts. Like other influenza A viruses, equine influenza virus undergoes variation through a process known as antigenic drift. The proteins on the surface of the virus particles gradually mutate, which is how they avoid recognition by the host immune system and it also allows the viruses to change their binding specificity facilitating the cross species transmission events seen with influenza A viruses. Influenza viruses can be divided into subtypes, based on antigenic analysis of the surface glycoproteins haemagglutinin (HA) and neuraminidase (NA) (Fig 1, Table 1). The HA protein is responsible for the binding of virus particles tothe surface of cells and the fusion between the virus and cell membranes. It is also a major target for the antibody response. Antibodies to HA will protect horses against influenza viruses of a similar antigenic type by binding to the HA protein and interfering with the attachment to the cell surface. Haemagglutinin is therefore an important component of vaccines. The NA glycoprotein is the other major surface protein and determinant of subtype and it acts by cleaving the progeny viruses from the surface of infected cells allowing them to disseminate and infect other cells. Antibodies against NA do not block infection, but they can inhibit the enzymatic activity of NA, which can decrease viral replication. Little is known about the importance of NA-specific immune response for protection against equine influenza.

Table 1. Host range of influenza subtypes
SubtypeInfluenza A virus natural host range
 H4X X  
 H7XX X 
 N3X X  
 N6X X  
 N7XX X 
Figure 1.

Electron micrograph of EIV (a) and a cartoon showing the structure of the influenza A virus (b). It has a segmented RNA genome consisting of 8 segments which are coated in protein (c).

Horses have been infected by 2 different subtypes, H7N7 and H3N8 (Table 1) (Sovinova et al. 1958; Waddell et al. 1963). All viruses isolated in the last 30 years from horses have belonged to the H3N8 subtype suggesting that the H7N7 viruses are no longer circulating the horse population. Equine influenza viruses appear to vary less than human influenza viruses at the antigenic level, which means that vaccines do not have to be updated every year to remain effective. However, there have been a number of outbreaks in vaccinated Thoroughbreds over the years (Daly et al. 2004a). In an attempt to ensure that vaccines remain up to date, several countries participate in surveillance schemes to collect data about viruses that are currently circulating. Viruses are isolated from nasal swabs taken from affected horses, they are then characterised at the genetic level by sequencing the HA glycoprotein gene and at the antigenic level by carrying out haemagglutination inhibition (HI) assays (Bryant et al. 2009b).

Phylogenetic analyses of the HA sequences enables family trees to be drawn to show the relationship between strains (Fig 2). Work partially funded by the HBLB showed that the equine influenza viruses diverged into 2 lineages, the European and American, in the late 1980s (Daly et al. 1996). Viruses from the 2 lineages were sufficiently different antigenically that a horse vaccinated with a virus from the American lineage would not be well protected against a virus from the European lineage (Daly et al. 2004b). More recently, the American lineage has evolved into 2 further sublineages, Florida sublineage Clade 1 and Clade 2 (Lai et al. 2001; OIE 2008). The HBLB supported surveillance programme has isolated both Florida Clade 1 and 2 viruses from the UK in 2007 to 2010 (Bryant et al. 2009b, A. Rash and A. Woodward, unpublished data) (Fig 2). Both of these virus groups are in circulation in other parts of the world and have caused large outbreaks of equine influenza in Australia (2007) (Callinan 2008), Japan (2007) (Yamanaka et al. 2008), Mongolia (2008) and India (2008–09) (Virmani et al. 2010). Due to the transport of horses all over the world for events and breeding, equine influenza virus in one continent is a risk to horses in other areas of the world. The outbreaks in Australia and Japan were caused by viruses similar to those circulating in the USA in previous years.

Figure 2.

Phylogenetic analysis of the HA1 nucleotide sequences encoded by equine influenza virus using PhyML version3.

Haemagglutination inhibition assays measure the ability of individual virus strains to be recognised by specific antisera and give a good indication of whether a virus has mutated significantly compared to other strains or not. A set of ferret antisera raised to 10 different viruses are used to compare the antigenic profiles of the field isolates. However, the data generated can be difficult to interpret. The technique of antigenic cartography has been used for human influenza to help visualise the antigenic relationship between virus strains (Smith et al. 2004). This technique has been applied to equine influenza virus data obtained through the surveillance programme. It has enabled virus strains to be plotted on a map, based on their HI activities, with strains that are closer together on the map being more closely related at the antigenic level. The greater the antigenic distance between viruses isolated in the field and the strains used in vaccines, the greater the chance that there will be a breakdown in protection offered by those vaccines (Mumford 2007).

Ultimately, protection studies in horses are carried out to test the ability of vaccines to protect against field strains of virus. Most studies carried out by vaccine manufacturers are set up to test protection shortly after completion of the vaccination course, when the horses' antibodies to influenza are at the highest level. Under these conditions, vaccines usually work very well. These studies are useful when looking at the possibility of protecting horses in the face of an outbreak, where horses at risk may be vaccinated. This is one of the approaches used to control the outbreak in Australia in 2007. Work sponsored by the HBLB showed vaccines available in the UK in 2007 were able to partially protect against the 2007 Australian outbreak virus, if it was to have arrived on our shores (Bryant et al. 2009a). However, experiments also need to be carried out specifically to study the ability of one virus strain to protect against another, to obtain the information needed to make good vaccine selection decisions. For human influenza viruses, which cannot be routinely tested in man for ethical reasons, mouse and ferret model systems are often used. One of the areas currently being investigated as part of the equine influenza surveillance programme is the ability of a mouse model to act as a substitute for experiments in horses.


As a result of the introduction of mandatory vaccination for racing Thoroughbreds in 1981, the racing industry of the UK committed to a long-term programme to monitor vaccine efficacy and to conduct ongoing surveillance of equine influenza virus. Surveillance is also carried out in other countries, but generally remains poorly funded. Within the UK this work is funded by the HBLB and is carried out by the Animal Health Trust, an OIE reference laboratory for equine influenza virus. Nasal swabs are collected by equine practitioners from horses suspected of influenza and are sent to the Trust for diagnostic screening. Virus is isolated from positive samples and characterised as described above. A sentinel practice scheme has been established, through which diagnostic testing can be carried out free of charge. Currently there are 55 equine practitioners registered for this scheme, each participant receives a sampling pack plus regular newsletters. There is also a dedicated website with further information about outbreaks and characterisation of viruses (http://www.equiflunet.org.uk). There are additional schemes in place, sponsored by vaccine manufacturers, including a text alert rapid notification scheme (‘Tell-tail, Merial') and an equine respiratory screening service (‘RespCheck, Intervet-Schering Plough Animal Health’). All these schemes aim to increase the number of samples submitted to the Animal Health Trust for diagnostic screening: the vaccine strain selection process can only work if there are viruses to review. The UK and Ireland have the most comprehensive surveillance systems in place, in other countries it has proven difficult to obtain funding for this type of research.

Collaborating laboratories share data and exchange reagents, all the information received is reviewed annually by the Expert Surveillance Panel (ESP). The ESP was created to improve the efficacy of equine influenza vaccines, by ensuring that vaccines contain epidemiologically relevant strains. A major challenge for the ESP is to decide when viruses circulating in the field have undergone significant antigenic drift that the vaccine strains need to be updated. The recommendations of the ESP are published by the OIE (World organisation for animal health) in their bulletins and can be downloaded from the OIE website at http://www.oie.int/publications-and-documentation/bulletins-online. They are made publically available and vaccine manufacturers are informed. It should be noted that there is no legal requirement for vaccine manufacturers to update their strains following a change to the OIE recommendations. The process is extremely costly and time consuming and it usually takes several years for a change in the recommendations to make it through to commercially available vaccines. We would urge veterinary practitioners to use vaccines with current recommended strains, wherever possible.

Mandatory vaccination was introduced by the Jockey Club for racing Thoroughbreds in 1981, in response to the extensive outbreak of equine influenza in 1979 that caused widespread disruption to the equine industry. This programme was effective for several years, then a further outbreak occurred in 1989 in which vaccinated horses were not protected. Since then, 3 major changes have been recommended to vaccine strains. The first formal recommendation to update strains was made in 1993 and referred to the need to replace out-of-date vaccine strains from 1979–81 with viruses isolated in 1989. In the mid 1990s it became clear that equine H3N8 viruses had diverged into 2 sublineages, designated European and American, which were antigenically distinct. In 1995 a recommendation was made that vaccines should contain a representative of both sublineages. No major breakdowns in vaccination occurred in the UK until 2003, when over 1000 horses were affected in Newmarket. Later the same year, a huge outbreak occurred in South Africa, where no vaccination programme was in place. In 2004 the ESP recommended that a further update was necessary for the American lineage viruses: the 1993/1994 viruses should be replaced by viruses antigenically similar to South Africa/4/2003. This decision was based on field infections in vaccinated horses and antigenic differences determined in HI assays using ferret sera. The most recent change in the recommendations were made in early 2010: there is no longer a requirement for the Eurasian lineage virus and the vaccines should contain a representative of both Florida sublineage Clades 1 (e.g. South Africa/4/2003, Lincoln/07) and 2 (Richmond/1/2007).

Unfortunately, updates to commercial vaccines generally take place following a major outbreak of equine influenza, on the scale of the Australian outbreak or of large scale vaccine breakdown. The long-term goal of the international surveillance schemes and ESP is to predict when breakdown is likely to occur and encourage update of strains before such an event. The financial constraints on vaccine manufacturers and the lack of demand by equine practitioners in the field for up to date vaccines remain problematic. The recent epidemics of equine influenza in a vaccinated horse population in Japan and a naïve population in Australia demonstrate the huge economic consequences of major outbreaks. These events highlight the importance of supporting schemes that contribute to better vaccines through appropriate selection of strains and monitoring that products meet the potency standards laid down by the OIE. The existence of a network of collaborating centres with experts in equine influenza around the world is crucial to provide a global service for management of such outbreaks. The recent pandemic of swine influenza in man shows how rapidly influenza can spread in a susceptible population. Unlike most other animals, high value competition horses are also transported around the world for events and breeding and are vulnerable to infection unless adequately protected.


The majority of the competition horse population is vaccinated against EIV in the UK. However, much of the remaining horse population is not vaccinated and can act as a virus reservoir. After vaccination, protection from virus infection has been shown to correlate with the similarity of the vaccine to the infecting virus strain (Yates and Mumford 2000). A significant mismatch of strains generates a situation where vaccines provide reasonable clinical protection so disease is mild and short lived, but infected animals shed large amounts of virus and fuel spread of infection.

The level of circulating antibodies against the viral HA has been shown to play an important role in protection against equine influenza virus (Mumford and Wood 1992). This surface glycoprotein is essential for virus entry into cells in the respiratory tract. The presence of specific antibodies that bind to HA in the airways and blood can neutralise viral particles and prevent infection of cells.

Historically, the majority of commercial vaccines have consisted of inactivated whole virus (Table 2). Influenza viruses are grown in embryonated hens' eggs or cell culture then inactivated chemically. This inactivated virus is often mixed with an adjuvant that improves the strength and duration of the immune response. The main advantage of these vaccines is that they are safe and will not replicate or cause disease in the horse. Protection induced by these inactivated EIV vaccines is based on the stimulation of a strong antibody response in the horse mainly directed against HA, but it requires a good match between vaccine and field strains of virus. These vaccines do not stimulate the cellular branch of the host immune system, which is thought to be an important component of the long lasting protection that results from natural infection. Unfortunately, the antibody response to equine influenza vaccines if fairly short lived and horses need to receive booster vaccinations regularly. In man, it is thought that cellular immune mechanisms play an important role in clearance of virus from the respiratory tract (McMichael et al. 1983). Natural infection with EIV has been shown to induce long-term immunity independent of circulating antibodies against HA in the horse (Hannant et al. 1988; Bryant et al. 2009a).

Table 2. Commercial vaccines available in the UK
Trade nameCompanyTechnologyAdjuvantVirus strainsSelected references
ProteqFluMerial Animal Health LtdPureVax -recombinant canarypoxCarbomerOhio/03 (H3N8) Newmarket/2/93 (H3N8)Edlund Toulemonde et al. (2005)
Equilis PrequenzaIntervet/Schering Plough animal healthInactivatedISCOM-matrixPrague/56 (H7N7) Newmarket/1/93 (H3N8) Newmarket/2/93 (H3N8)Heldens et al. (2010)
Equip-FPfizer LtdInactivatedSelf adjuvanted (ISCOM)Newmarket/77 (H7N7) Borlänge/91 (H3N8) Kentucky/98 (H3N8)Paillot et al. (2008)
Duvaxyn IE PlusElanco Animal HealthInactivatedCarbomer & Aluminium hydroxideNewmarket/1/93 (H3N8) Suffolk/89 (H3N8) Prague/56 (H7N7)Paillot et al. (2010)

In the last 15 years, a new generation of vaccines has been designed to stimulate both antibody and cellular immune responses that mimic the immunity induced by natural infection with equine influenza virus (Table 2) (for review see Hannant et al. 1988; Paillot et al. 2006a). Subunit vaccines against equine influenza virus contain purified virus proteins rather than whole virus. The 2 main types of vaccines are either the immune-stimulating complexes (ISCOM)-based vaccines or ISCOMATRIX vaccines. ISCOM particles are spontaneously formed cage-like structures resulting from the combination of viral proteins with cholesterol, phospholipids and Quillaja saponins. ISCOMATRIX vaccines are similar to the ISCOM vaccines except the viral proteins are not bound within the cage structure of the ISCOM (Paillot et al. 2008). Another new development has been the use of live-vectored vaccines in horses. DNA encoding the HA protein from influenza viruses has been inserted into a poxvirus, in this case Canarypox. This virus is live and infectious, but will not replicate or cause disease in the horse. However, HA proteins expressed by cells infected with the recombinant canarypox vaccine vector will result in the induction of antibodies and priming of cell mediated immune responses (Paillot et al. 2006b). The final class of vaccines is the live-attenuated virus vaccines, which consist of a temperature-sensitive influenza virus. This virus multiplies efficiently in the cooler environment of the upper respiratory tract where immune responses are induced, but does not replicate in the warmer environment of the lower respiratory tract in horses. It is therefore nonpathogenic and cannot cause clinical signs of disease, but effectively stimulates a strong immune response. This vaccine is licensed in the USA but is currently not available in the UK.

The duration of immunity following vaccination may be variable, dependent upon the specific vaccine used and response of the individual horse. There are some horses that do not develop a normal antibody response to vaccination, despite the correct vaccination protocol being followed. These represent a very small percentage of horses and the phenomenon is not yet fully understood; however, it has been associated with the age of the horse (Horohov et al. 2010). It has been suggested in the past that vaccination against equine influenza can make horses sick and days in training can be lost as a result. Our experience suggests that some inactivated virus vaccines had documented problems, however the vaccines currently available have minimal associated side effects and are commonly used in actively training horses. The international travel of horses for racing relies on quarantine and vaccination to control infection and the equine industry requires products which minimise virus shedding. Thus there are significant benefits to using the most effective vaccines, containing up to date strains.


Historically, the gold standard laboratory test for equine influenza virus has been virus isolation from nasal swabs in hens' eggs followed by a haemagglutination assay to detect virus. Although this method is very sensitive for some virus isolates, not all strains grow well in eggs and it may take up to 5 passages to obtain significant amounts of live virus. The condition of the diagnostic samples is also critical, as only live virus will grow. Overall, this method is very time consuming for a provisional diagnosis on which decisions such as vaccination, quarantine and movement bans may be based. More recently, assays have been developed that rely on detecting viral proteins, rather than live virus (Table 3). These assays generally work by using antibodies specific for influenza virus proteins that bind to the virus, then adding a chemical reagent that generates a visible colour change when binding has occurred. The diagnostic laboratory at the AHT uses an ELISA assay to detect the virus nucleoprotein, a highly conserved protein that is stable in nasal swab extracts that have been sent through the post (Cook et al. 1988). Some of the commercially available rapid assay kits can give a result in as little as 15 min. However, these diagnostic kits are usually a lot less sensitive than virus isolation in eggs (Chambers et al. 1994; McCabe et al. 2006). This is in contrast to the recently developed real time PCR methods that have become increasingly popular in diagnostics following the outbreak in Australia, which are extremely sensitive. It is now compulsory to test horses by this method pre- and post export to this country. Various laboratories have designed protocols based on this technology, that involves the amplification of a very small region of the viral genetic material. Through each cycle of amplification the number of DNA copies can be quantified, so the initial number of viral genomes present in the sample being tested can be estimated. This method has proved as sensitive as isolation in eggs; however, it can also detect traces of dead virus. When trying to diagnose equine influenza from a nasal swab submitted from a field case, this can be an advantage. When trying to determine whether a horse in quarantine is infectious or not, this can be a disadvantage. Efforts are currently being made to standardise all the different protocols internationally and to agree guidelines on the interpretation of results.

Table 3. Diagnostic tests to detect equine influenza virus infections
EIV diagnostic testsSelected references
Capture ELISACook et al. (1988)
Rapid diagnostic kitsChambers et al. (1994) McCabe et al. (2006)
RT-PCRBryant et al. (2009b)
Quantitative RT-PCRBryant et al. (2009a)
ImmunofluorescenceAnestad and Maagaard (1990)
Cell isolationMeguro et al. (1979)
Haemagglutination inhibition test (HI)Hirst (1942)
Single radial haemolysis assay (SRH)Schild et al. (1975)

Serological techniques can also be used to determine if a horse has been infected with equine influenza virus. Antibodies made in response to infection can be detected at high levels from approximately 2 weeks to about 3 months post infection, after which they quickly decline in most horses. Haemagglutination inhibition assays using a panel of reference viruses can be used to determine the subtype and lineage to which it belongs. If diagnosis is based only on serology, it can be difficult to determine whether a horse has been vaccinated or has had disease, particularly if only one blood sample is sent in for testing. The presence of H7N7 equine influenza virus strains in vaccines is no longer recommended, but several vaccines still contain an H7N7 strain as a useful marker. As equine H7N7 is not currently circulating, the presence of H7N7 specific antibodies is a strong indication that antibodies have been induced by vaccination rather than infection. Paired sera collected at an interval of 2–3 weeks are more useful and an increasing titre of equine influenza antibodies is indicative of infection if the animal has not been recently vaccinated.

Mathematical modelling

In recent years, the HBLB has supported work that has applied mathematical modelling techniques to equine influenza infections. Modelling has been used to look at the importance of mismatches between vaccine strains and circulating field viruses (Park et al. 2004). The model was based on a typical population of a Thoroughbred race training yard. Such modelling techniques take into account a variety of factors, such as size of population and age distribution and attempt to account for chance events such as the entry of a new horse onto the yard. The modelling results indicated that mismatches between vaccine and outbreak strains do make a difference to the likelihood that a population will be protected. Mathematical modelling was also used to analyse the data from the extensive outbreak of equine influenza in Newmarket during 2003 (Barquero et al. 2007). This outbreak was unusual in that there was a reasonable match between the vaccine and outbreak strain. Modelling experiments concluded that, unusually, older Thoroughbreds that had been vaccinated multiple times were more susceptible to infection, male horses were more susceptible than female and that type of vaccine had an effect. When an aluminium hydroxide adjuvanted vaccine containing antigens from Prague/56, Newmarket/1/93 and Newmarket/2/93 were administered first or last in a vaccine course it was associated with an increased risk of infection compared to other vaccine types. The biological reason for this difference was not clear. More recently, modelling techniques have been applied to test the effects of horses that are poor responders during a large outbreak, affecting multiple yards and also the use of vaccination in the face of an outbreak (Baguelin et al. 2010). Interestingly, the results suggest that the majority of equine influenza outbreaks are very limited and fail to spread. Vaccination in the face of an outbreak can have a significant beneficial effect for an individual trainer, but the effect is enhanced if other trainers also revaccinate.


Influenza A virus infection is initiated by virus attachment to sialic acid-containing cell-surface receptors. These surface receptors differ in their distribution throughout the tissues of their hosts and are an important influence on virus host range and tissue tropism (Suzuki et al. 2000). Equine influenza virus infection is restricted to the respiratory tract where it predominantly binds to α2-3 linked sialic acid. These receptors have been identified on ciliated epithelial cells in the nasal mucosa, trachea and in the bronchus of the horse (Muranaka et al. 2010). Cilia that have separated from epithelial cells because of damage caused by the virus have been detected in samples collected from naturally and experimentally infected ponies. Virus antigen has been detected in cell samples collected from the nasopharynx, trachea, bronchus and alveoli by bronchoalveolar lavage (Sutton et al. 1997). The initial nonspecific immune response triggered by epithelial cell infection leads to the production of interferons and proinflammatory cytokines released into the respiratory tract and help fight the infection. These cytokines can have a direct effect on clinical signs of infection by causing fever, increased mucous production and inflammation in the respiratory tract and their levels often correlate with the viral titre. Some virus strains seem to be more pathogenic than others, indicated by more severe or more prolonged clinical signs. Differences in pathogenicity between viruses are difficult to quantify, but may be linked to the induction of different levels of proinflammatory cytokines, namely IL-6 and IFN-α that both have a key role in both symptom formation and host defence (Wattrang et al. 2003). The induction of cytokines by influenza viruses is linked in part to the NS1 protein, which interferes with the induction of type 1 interferons. Different strains appear to induce different levels of cytokines in the horse (Wattrang et al. 2003; Quinlivan et al. 2007). This feature has been exploited for the generation of mutated viruses that are less pathogenic. By deleting a region of the NS1, new recombinant strains have been made that cause less severe clinical signs in horses and have potential use as live attenuated vaccines (Chambers et al. 2009).

Cross species transmission

In 2004, the first reports appeared from the USA of cross species transmission of equine influenza virus to racing greyhounds (Crawford et al. 2005). In some cases, infection proved fatal and the initial mortality rate was estimated at around 10%. Since then, the virus has adapted to its new host, has spread to most states and the disease is typically mild. However, it has become established in rescue centres where it causes problems in dogs in poor condition. Within the UK, there have been 2 separate outbreaks of equine influenza in dogs (Daly et al. 2008, Newton et al. 2007). Both outbreaks were in quarry hounds, where animals were in close contact with horses. Following the outbreak of equine influenza in Australia in 2007, evidence of sporadic cases in dogs were recently reported (Kirkland et al. 2010). Transmission directly from a horse to a dog has been confirmed in an experimental setting in by co-housing an infected horse with a beagle in a stable (Yamanaka et al. 2009). It is not yet known whether virus can be transmitted from an infected dog back to a horse. However, the canine influenza viruses circulating in the USA have several mutations compared with equine viruses and it is thought that they are fully adapted to their new host.

In terms of cross species transmission of influenza into horses since the initial case back in the 1960s, there have been 2 documented outbreaks (Waddell et al. 1963). In the late 1980s, a new H3N8 sutbtype virus transmitted from birds to horses in China. This virus had a high mortality rate, killing around 10% of the animals infected (Webster and Thomas 1993). Fortunately this virus did not spread to other countries. More recently, the transmission of a highly pathogenic H5N1 ‘bird flu’ virus into donkeys has been reported in Egypt, where it caused only moderate disease (Adbdel-Moneim et al. 2010). These events highlight the potential threat of other sources of influenza and, although such cross species transmission events are extremely rare, we should be vigilant for other influenza subtypes.


The HBLB funded equine surveillance programme contributes the majority of the data available for vaccine strain selection and in doing so helps reduce the impact of equine influenza virus on the horse population for the benefit of both the industry and pleasure owners alike. With large epidemics of divergent equine influenza viruses occurring elsewhere in the world and the ability of influenza to cross the species barrier, surveillance has never been so important in order to keep the equine population in the UK as well protected as possible. In the future we plan to continue the surveillance programme with greater emphasis on obtaining virus samples from further afield to pre-empt any potential incursion of novel viruses from abroad. Research will also continue into the factors affecting the pathogenicity and spread of equine influenza viruses, including the development of in vitro and small animal models for the detailed study of these important viruses.

Conflicts of interest

HBLB's Veterinary Advisory Committee commissioned and sponsored this article as part of a series summarising progress made in areas relating to their priorities for research funding. The authors and other workers at his/her institute hold current and previous research grants funded by the HBLB. The authors are members of the team responsible for the Equine Influenza Programme funded by the HBLB.

EVJ is delighted to publish HBLB's Advances in Equine Veterinary Science and Practice Review Series in recognition of the major contribution that HBLB research and educational funding has made to the health and welfare of the Thoroughbred.