Influenza InfectionsJ. M. Daly1 and A. Cullinane2
Classification of Equine Influenza VirusEquine influenza viruses are classified as type A influenza viruses, which also include viruses infecting man, birds and swine. Type A influenza viruses are divided into subtypes according to the antigenic characteristics of the surface antigens, haemagglutinin (HA) and neuraminidase (NA). With the exception of the recently discovered H17N10 bat influenza virus , all the subtypes (combinations of H1 to H16 and N1 to N9) have been isolated from aquatic birds, which act as a natural reservoir for influenza A viruses. Two subtypes have established in equids (H7N7 and H3N8, previously known as equi-1 and equi-2, respectively).
EpidemiologyInfluenza virus was first isolated from horses in Czechoslovakia in 1956 ; this was the prototype strain for the H7N7 subtype of equine influenza virus A/equine/Prague/56 (H7N7). There has been no convincing evidence for circulation of H7N7 viruses in horses for several decades.
Virus of the H3N8 subtype was first isolated in Miami in 1963 . Antigenic variants can give rise to large-scale disease epidemics in regions where the virus is endemic, such as Europe and North America. Only New Zealand and Iceland are known to have remained entirely free from equine influenza. Some of the most devastating epidemics of equine influenza have occurred when the virus has been introduced into areas previously free of the disease. Outbreaks in South Africa (1986 and again in 2003/4), India (1987), Hong Kong (1992), and Australia (2007) highlighted the ease with which equine influenza can be introduced into susceptible populations as a result of international movement of horses [4-9]. The outbreak in Australia in 2007 particularly emphasised the potential impact of introduction of equine influenza virus into an unvaccinated population, with around 76,000 horses affected and an estimated expenditure of 360 million Australian dollars to bring the outbreak under control .
A novel strain of equine influenza caused an epidemic in China in 1989 with an overall morbidity rate of 80% and a mortality rate of 20%. The prototype virus from this epidemic (A/equine/Jilin/89) was shown to be more closely related to influenza viruses infecting aquatic birds than to contemporary equine influenza viruses, demonstrating that the epidemic was the result of emergence of virus from the avian influenza reservoir . A related avian-like virus, which had lost its ability to replicate in birds, affected a few hundred horses in 1990, but seems to have neither persisted in nor spread beyond China .
Influenza A viruses are unpredictable by their very nature and unlikely to ever be fully eradicated due to the reservoir of viruses circulating in wild aquatic birds. Recently, donkeys in Egypt joined the relatively long list of mammalian species to have been naturally infected with highly pathogenic avian influenza virus of the H5N1 subtype . It was thought that there was no interchange of influenza virus infections between horses and other mammalian species. However, interspecies transmission of equine influenza virus to dogs has occurred apparently independently in the USA, UK and Australia [14-16]. Furthermore, the H3N8 virus has become endemic, with dog-to-dog transmission, in North America. During surveillance of pigs in China, two H3N8 isolates closely related to equine viruses circulating in the 1990s were isolated in 2005 and 2006 . Although human volunteers have been experimentally infected with equine influenza virus , there is currently little evidence of zoonotic infection of people with equine influenza.
Clinical Signs and PathologyEquine influenza is contracted by inhalation. The virus infects the ciliated epithelial cells of the upper and lower airways and can cause deciliation of large areas of the respiratory tract within 4 to 6 days (Fig. 1a, Fig. 1b, Fig. 1c). As a result, the mucociliary clearance mechanism is compromised and tracheal clearance rates may be reduced for up to 32 days following infection .
Figure 1a. Effects of equine influenza virus infection on ciliated respiratory epithelium; (a) healthy epithelium. To view click on figure
Figure 1b. Effects of equine influenza virus infection on ciliated respiratory epithelium; (b) 3 days post infection. To view click on figure
Figure 1c. Effects of equine influenza virus infection on ciliated respiratory epithelium; (c) 6 days post infection, large areas of deciliation observed. To view click on figureClassic clinical signs of influenza infection in previously naïve animals are easily recognisable, although some strains of equine influenza are less virulent than others . In addition, the amount of virus to which the horse is exposed influences the severity of the disease. Influenza is characterised by a short incubation period of one to 3 days and rapid spread among susceptible animals. The rapidity with which influenza spreads is a key observation for differentiating influenza from other infectious respiratory diseases. The first sign is an elevation of body temperature (up to 41°C), which is usually followed by a deep dry cough that releases large quantities of virus into the atmosphere. This is often accompanied by a serous nasal discharge (Fig. 2), which may become mucopurulent. Other clinical signs that may be observed include: depression, myalgia, inappetance, enlarged submandibular lymph nodes, and oedema of the legs and scrotum. Previously healthy adult horses usually recover from uncomplicated influenza within 10 days, although coughing may persist for longer. If secondary bacterial infection occurs it can prolong the recovery period. On rare occasions more severe signs are observed, sometimes with fatal outcome. Neurological signs were observed in two influenza-infected horses during the 2003 outbreak in the UK , and broncho-interstitial pneumonia was observed in young foals without maternal antibodies during the 2007 outbreak in Australia .
Figure 2. Serous nasal discharge as seen with infection with equine influenza. To view click on figureThe severity of the disease varies with the immune status of the horse; in animals partially immune as a result of vaccination or previous infection there may be little or no coughing or pyrexia [23,24]. Whereas spread of infection throughout a group of naïve animals is generally rapid, outbreaks have been described in which the infection circulated sub-clinically in vaccinated horses for 18 days before inducing recognisable clinical signs . There have also been positive diagnoses of a few animals infected with equine influenza that have occurred without spread to other animals within Thoroughbred racehorse training yards . In well-vaccinated Thoroughbreds, loss of performance may be the only indication that a horse is infected .
DiagnosisVarious agents may cause outbreaks of infectious respiratory disease in horses including: equine herpes, rhino, adeno, and arteritis viruses, Streptococcus equi, S. zooepidemicus, or S. pneumoniae. Therefore, a presumptive diagnosis of influenza based on clinical signs should be confirmed by laboratory testing. Traditionally, laboratory confirmation of a clinical diagnosis has been by isolation of virus from nasopharyngeal swabs or retrospective serology, both processes that can take several days if not weeks. Nowadays, a wide range of laboratory tests are available that allow more rapid diagnosis of influenza infection.
Nasopharyngeal swabs for virus isolation or detection should be taken as promptly as possible. Results of experimental challenge studies suggest that peak viral titres occur on the second or third day after infection, and duration of viral shedding is usually not more than 4 or 5 days in naïve animals . Nasal swab samples are taken by passing a swab as far as possible into the horse’s nasopharynx via the ventral meatus to absorb respiratory secretions. Swabs should be transferred immediately to a container with virus transport medium, which contains antibiotics, and kept cool to maintain viability of the virus. Most equine diagnostic laboratories will supply appropriate swabs and transport medium (Fig. 3). Virus is unlikely to survive if dry swabs are taken and there is increased chance of contamination if bacterial transport medium is used. Most laboratories only attempt virus isolation from samples that have tested positive in a rapid diagnostic test. Virus isolation is essential to surveillance programmes and provides isolates to update vaccine strains when this becomes necessary to maintain vaccine effectiveness (Fig. 4). Where there are low levels of virus shedding, such as in vaccinated horses, virus isolation may prove difficult and some strains are more difficult to isolate than others.
Figure 3. Equipment required for nasopharyngeal swabbing of the horse; sterile swab consisting of gauze attached to a length of soft stainless steel wire, tube containing transport medium into which the gauze end of the swab is placed and the wire trimmed off, insulated container with refrigerant pack and form for submission to diagnostic laboratory. To view click on figure
Figure 4. Inoculation of embryonated hens’ eggs with equine influenza virus. To view click on figureVirus isolation has been largely replaced by rapid diagnostic tests that detect the presence of viral antigens (enzyme-linked immunosorbent assays, ELISAs) or viral nucleic acid (reverse-transcription and polymerase chain reaction, RT-PCR) in nasal swab extracts. RT-PCR is the test of choice in most laboratories as it is highly sensitive and can provide a result in a matter of hours [28-30]. An equine influenza-specific ELISA to detect viral nucleoprotein has been in use at the Animal Health Trust, Newmarket, UK since 1989 [31,32]. There are now commercial influenza detection kits available for the diagnosis of human influenza that cross-react with equine influenza. Although less sensitive than equine-specific assays, these kits are easy to use by staff with little laboratory training and provide a rapid result [33-36].
Influenza infection can also be diagnosed by comparison of the results of serological testing of an acute serum sample taken as soon as possible after the onset of clinical signs and a convalescent serum sample taken around 2 weeks later. There are two serological assays – the haemagglutination inhibition (HI) test (Fig. 5) and the single-radial haemolysis (SRH) test (Fig. 6). Both assays rely on measurement of antibodies to haemagglutinin, the surface glycoprotein responsible for viral attachment and entry into host cells and the primary target of virus neutralising antibodies. The SRH test is more complicated and time-consuming to perform than the HI test, but has the advantage of being easier to standardise between laboratories and is therefore particularly useful for vaccine efficacy and epidemiological studies. Serological detection of influenza virus infection is rarely useful in immediate clinical management. However, it is a useful surveillance tool and may be essential to establish a diagnosis where virus detection has not been possible (virus may only be shed for 1 or 2 days and at lower levels in partially immune animals). ELISA tests that detect antibodies against the viral nucleoprotein have recently become available. They are less sensitive than HI and SRH for the detection of influenza in populations where the virus is endemic, but were extremely useful in the control and eradication of equine influenza in Australia . Most recently, a sensitive and quantitative assay to measure neutralising antibodies was developed using an equine influenza pseudotyped lentivirus .
Figure 5. Haemagglutination inhibition (HI) assay showing 6 serum samples with titres of 32 (S1, S2 and S3), 16 (S4) and <8 and="" back-titrated="" class="red" confirm="" correct="" dose="" haemagglutinating="" is="" s6="" span="" standard="" that="" the="" to="" units="" used.="" virus="">To view click on figure8>
Figure 6. Single radial haemolysis plate showing clear zones of haemolysis around wells for 9 out of 16 serum samples and a positive control sample in well 16. To view click on figure
VaccinationTraditional equine influenza vaccines consist of representative virus strains either as inactivated whole virus or protein subunits and are dependent on induction of antibodies to the surface glycoproteins, in particular haemagglutinin. Levels of antibody required for protection of horses have been determined through vaccination and challenge studies and from field data [39,40]. The vaccine-induced antibody response to HA is remarkably short-lived in horses, therefore, adjuvants such as aluminium hydroxide, carbomer or saponins are normally included to enhance the height and duration of the immune response to whole virus or subunit [41,42].
Live-attenuated and canarypox-vectored vaccines are now also available against equine influenza in some countries. These vaccines are intended to mimic the immune response to natural infection more closely than the traditional vaccines. However, both second-generation vaccines such as immune-stimulating complex (ISCOM)-based and canarypox recombinant vaccines on the one hand, and a more traditional whole inactivated vaccine on the other, have been shown to elicit a cell-mediated immune response [43-46].
Vaccination against equine influenza is mandatory for racehorses and competition horses in several countries and it is important to comply with the relevant regulations. This does not, however, guarantee protection of individual horses, although vaccinated horses typically have milder disease than non-vaccinated horses, are less likely to develop secondary bacterial infections and can thus return to work sooner. A few horses respond poorly to vaccination and have persistently low antibody titres; such horses are often instrumental in starting an influenza outbreak within a premise . Serological monitoring of antibody levels can be useful to determine whether additional booster vaccinations are required . Equine influenza outbreaks are frequently associated with sales or race meets where horses from different regions congregate and mix. It may therefore be advantageous to time additional booster vaccinations to be given prior to such events.
It has been demonstrated that inclusion of an additional booster vaccination to close the window of susceptibility between the second and third vaccinations recommended by the vaccine manufacturers is of benefit to young horses . Furthermore, it is advisable to vaccinate young horses, particularly racehorses and other competition horses, at 4- to 6-month intervals for several years after their primary course of vaccinations. An annual booster will usually suffice for older horses such as show jumpers and brood mares that have been vaccinated regularly since they were foals.
Brood mares should be vaccinated in the later stages of pregnancy, but not later than 2 weeks prior to foaling, to ensure a good supply of colostral antibodies for the foal. For most vaccines, vaccination in the presence of maternal antibodies is not recommended as they interfere with the response to vaccination and may even lead to the induction of serotolerance [49-51]. It is therefore advisable to monitor the antibody titres of foals and start the vaccination programme when maternal antibody titres have declined to negligible levels, usually at around 6 months of age (foals from non-immunised dams should be vaccinated earlier). It has been suggested that vaccination of foals with the canarypox-vectored recombinant vaccine effectively primes the immune system despite the presence of maternally-derived antibodies .
Some existing vaccines are available as combination vaccines also containing tetanus or herpes antigens. The immune response elicited by tetanus toxoid is much more durable than that induced by influenza antigen and it is inadvisable to administer tetanus toxoid more frequently than the vaccine manufacturers recommend. In an intensive influenza vaccination programme, vaccines containing influenza only should be routinely administered.
Vaccine Strain SelectionIt has been demonstrated in vaccination and challenge studies and in the field that the ability of a vaccine to reduce virus shedding is directly correlated with the antigenic relatedness of the vaccine strain and the challenge virus [53,54]. A European rapid licensing system for influenza vaccines containing updated strains has been developed on the basis that accurate standardisation of in vitro vaccine potency tests allows a reduction in animal testing of the final product . It is nonetheless a major undertaking to update vaccine strains. An Expert Surveillance Panel was instigated at a meeting of World Health Organisation and OIE (World Organization for Animal Health) experts in 1995 . Human influenza vaccines are reviewed on an annual basis and in most years at least one of the 3 components is updated. The equine influenza Expert Surveillance Panel similarly reviews available data annually and reports whether changes to vaccination composition are required. However, antigenic drift occurs at a much lower rate in equine influenza viruses than in human influenza viruses , therefore equine vaccines do not need to be updated so frequently. The initial recommendation of the Expert Surveillance Panel was that vaccines should contain one H3N8 virus representative of the "American-like" lineage and a representative of the "European-like" lineage . Prior to 1993, it was thought that equine influenza H3N8 viruses evolved in a single lineage . However, in 1993 viruses were isolated that revealed divergent evolution on the American and European continents ; the American and European lineages (Fig. 7). In recent years, there has been little evidence of European lineage viruses in circulation whereas the American lineage has continued to evolve with the predominance of a "Florida sublineage", which has further diverged into two distinct "clades" [59,60]. The latest recommendations to have been published are that vaccines should contain both clade 1 and clade 2 viruses of the Florida sublineage but that it is not necessary to include either an H7N7 strain or a "Eurasian lineage" strain .
Figure 7. Phylogenetic tree of the haemagglutinin (HA1 portion) of H3N8 equine influenza viruses. Horizontal distance is proportional to the number of nucleotide differences to join nodes and viruses. Vertical lines are for spacing branches and labels. To view click on figure
Management of an Influenza OutbreakTreatments for equine influenza are aimed at ameliorating clinical signs; antiviral agents have only been used experimentally [62-66]. Rest is essential to reduce the severity of clinical signs and the recovery period, and minimise chronic sequelae. Ill animals should be provided with plenty of fresh water. Electrolytes can be provided, in a second bucket of water . A non-steroidal anti-inflammatory drug (NSAID) can be used in cases with high fever to reduce pulmonary inflammation and in pregnant mares to reduce the risk of abortion. Clinical signs should be monitored daily and if secondary bacterial infection is suspected (indicated by coarse crackles sounds in the lung and prolonged or rising fever), antibiotic therapy should be initiated, and samples obtained for bacterial culture and antibiotic sensitivity testing.
Immediate and effective quarantine of infected individuals is the best way to minimise spread of influenza virus within premises. In order for quarantine measures to be effective, infected individuals must be rapidly identified and isolated. This can be achieved by regularly taking temperatures and / or samples for rapid diagnostic tests. Similarly, restricting movement from premises can limit spread of the virus further afield. In Australia during the 2007 outbreak of equine influenza outbreak, movement restrictions were put in place within and between four different zones according to the level of infection already observed within each zone . Precautions should be taken against the risk of spread by personnel or shared equipment (including vehicles used for transportation); spread of virus by fomites or people was implicated in both outbreaks of equine influenza in South Africa [7,8].
Vaccination in the face of an outbreak has also proven to be effective [10,68,69]. However, if horses are vaccinated in the face of infection it can be difficult, using HI and SRH assays, to determine whether any increase in antibody levels is due to vaccination or infection. In Australia, the canarypox-vectored vaccine was used as part of the effort to contain the outbreak in 2007. This vaccine expresses only the HA protein of equine influenza virus, therefore using an ELISA to detect antibodies to nucleoprotein (present in infectious virus) enables differentiation of infected from vaccinated animals (the so-called DIVA strategy).