Hendra virus was first described in 1994 after an outbreak of acute respiratory disease in a Thoroughbred training stable in Australia in which horses and one person were fatally infected. Sporadic cases continue to occur in eastern Australia, typically presenting as an acute febrile illness and rapidly progressing with variable system involvement, notably acute respiratory and/or severe neurologic disease. Fruit bats (suborder Megachiroptera) are the natural reservoir of the virus. Hendra virus is classified as a biosafety Level 4 agent (defined as posing a high risk of life-threatening disease in people), and the use of safe work practices and personal protective equipment is essential to manage the risk of human exposure. The earlier names of equine morbillivirus and acute equine respiratory syndrome are no longer appropriate.
Etiology and Pathogenesis
Hendra virus is a large, pleomorphic enveloped RNA virus. Although initially considered to be more closely related to members of the genus Morbillivirus than to other genera in the family Paramyxoviridae, subsequent studies showed limited sequence homology with respiroviruses, morbilliviruses, and rubuloviruses and negligible immunologic cross-reactivity with other paramyxoviruses. Hendra virus is genetically and antigenically closely related to Nipah virus (see Nipah Virus Infection), with which it shares >90% amino acid homology. Both viruses have been classified in a new genus, Henipavirus, in the subfamily Paramyxovirinae.
It is increasingly evident that Hendra virus strain variation is minimal and that clinical presentation and pathology more likely vary with the route of infection. Historically, interstitial pneumonia of variable severity was the principal finding in naturally infected horses and in experimentally infected horses exposed by the respiratory or parenteral routes. Hendra virus has a specific tropism for vascular tissues, regardless of route of challenge. In early infection, the vascular lesions may include edema and hemorrhage of vessel walls, fibrinoid degeneration with pyknotic nuclei in endothelial and tunica media cells, and numerous giant cells (syncytia) in the endothelium and sometimes the tunica media of affected vessels (both venules and arterioles). The virus becomes more widely distributed in various tissues throughout the body as infection progresses, presumably as a result of a leukocyte-associated viremia. Virus has been demonstrated in the vascular endothelium of subarachnoid and cerebral vessels and in the vasculature of the renal glomerulus and pelvis, lamina propria of the stomach, spleen, various lymph nodes, and myocardium. When respiratory disease is present, there is progressive destruction of alveolar walls, with the appearance of alveolar and intravascular macrophages. In addition to its vascular tropism, Hendra virus can also be neurotropic, causing neuronal necrosis and focal gliosis. A feature of one outbreak at an equine veterinary clinic in Australia in 2008 was severe neurologic disease and an absence of respiratory disease. Thus, Hendra virus should no longer be regarded as causing predominantly respiratory disease in horses.
Epidemiology and Transmission
Naturally occurring disease caused by Hendra virus has been reported only in horses and people. Experimentally, disease has been produced in cats, hamsters, ferrets, monkeys, pigs, and guinea pigs, but not in mice, rats, rabbits, chickens, or dogs. The clinical response and pathologic findings in cats are very similar to those seen in horses. Hendra virus infection and disease in horses has only been reported in Australia, and events are sporadic and infrequent, with 14 events recorded between 1994 and 2010. Most of these were single horse events. Thus, Hendra virus appears to have limited infectivity, and under field conditions, transmission between infected and noninfected horses occurs infrequently. However, the frequency of Hendra virus infection in horses has increased since 2011, with 18 incidents in 2011, 8 in 2012, 8 in 2013, and 3 in 2014 (up to October), with extended geographic locations (between far north Queensland to north New South Wales, Australia). In July 2011, a dog on a property with horses infected with Hendra virus (in Queensland) was identified as seropositive without any clinical signs. In July 2013, a dog on a property (in New South Wales) with Hendra virus infection in a horse was confirmed to be infected with the same virus. Further research is underway to investigate whether these increased incidents in horses were because of greater public awareness in reporting the disease or whether environmental or ecologic factors triggered this increased number of cases and extended geographic occurrence.
Experimentally, attempted transmission from virus-infected horses to in-contact horses or cats has been unsuccessful. Nonetheless, the possibility of respiratory transmission cannot be excluded. The frothy nasal discharge (originating from the lungs) sometimes observed terminally in naturally affected horses could plausibly provide a source of virus for aerosol transmission. Hendra virus has been found in the urine, blood, and nasal and oral secretions of naturally infected horses and cats. Based on available field and laboratory data, infection of people or animals appears to require direct contact with virus-infective secretions (lung exudates), excretions (urine), body fluids, or tissues. Although Hendra virus appears to have limited infectivity, the case fatality rate in individuals that become infected is high: 75% in horses, 57% in people.
Available epidemiologic, serologic, and virologic evidence implicates fruit bats as the natural reservoir of Hendra virus. Serologic surveys have revealed a high prevalence of neutralizing antibodies in wild-caught fruit bats (Pteropus spp) in Australia and Papua New Guinea. The geographic distribution of the virus in fruit bats appears to be limited to Australia and Papua New Guinea, although a transition of Hendra-like to Nipah-like viruses may occur beyond Australia. Infection in fruit bats (either natural or experimental) causes no evident disease. There is field and experimental evidence of vertical transmission, with isolates recovered from the uterine fluid and fetal tissues of a grey-headed flying fox (P poliocephalus) and a black flying fox (P alecto). The infrequent occurrence and sporadic nature of equine cases suggest that exposure of horses to Hendra virus is, at least in part, a chance event. The modes of transmission between bats, and from bats to horses, are uncertain, as are factors that may facilitate spillover. Hendra virus has been identified in the birthing fluids, placental material, aborted pups, and urine of naturally infected fruit bats and in the urine of experimentally infected fruit bats. The related Nipah virus has been detected in bat urine and on fruit partially eaten by bats. Horses are hypothesized to be infected through contact with food or water contaminated with material from infected fruit bats, but the definitive mechanism remains to be determined.
Because of its affinity for endothelial cells, Hendra virus can cause a range of clinical signs in horses. The predominant clinical presentation may depend on which organ system sustains the most severe or compromising endothelial damage.
Hendra virus infection should be considered when there is acute-onset fever and rapid progression to death, possibly associated with either severe respiratory or neurologic signs; however, the absence of these should not preclude consideration of Hendra virus. Infection is not always fatal, with 25% of known cases having recovered from clinical disease.
Clinical signs that should prompt a veterinarian to consider Hendra virus infection include acute onset of illness, fever, and rapid deterioration. Respiratory signs can include pulmonary edema and congestion, respiratory distress (increased respiratory rate), and terminal nasal discharge, which may be clear initially and progress to stable white or blood-stained froth. Neurologic signs can include “wobbly gait” progressing to ataxia, altered consciousness (apparent loss of vision in one or both eyes, aimless walking in a dazed state), head tilt, circling, muscle twitching (myoclonic spasms have been seen in acutely ill and recovered horses), urinary incontinence, recumbency with inability to rise, terminal weakness, ataxia, and collapse. Other clinical signs may include depression, highly increased heart rate, facial edema, muscle trembling, anorexia, congestion of oral mucous membranes, colic-like symptoms (generally quiet abdominal sounds on auscultation of the abdomen in preterminal cases), and stranguria in both males and females. Proximity to fruit bat roosts or feeding sites should increase the index of suspicion.
Where horses are paddocked, Hendra virus infection is more likely to manifest as a single sick or dead horse than as multiple cases. Most paddock infections have involved a single fatally infected horse with no transmission to in-contact companion horses. However, on several occasions, one or more companion horses have become infected after close contact with the index case before or at the time of death. Where horses are stabled, it appears that Hendra virus has the potential to spread either through close direct contact with infectious body fluids, or through indirect transmission via contaminated fomites, including inadvertent human-assisted transfer. Hendra virus infections in horse stables to date have resulted in multiple horses becoming infected, which appear to have arisen from a horse infected in a paddock or outside yard being brought into the stable.
The presence of large endothelial syncytial cells on histopathology is characteristic of Hendra virus infection. Although most prominent in pulmonary capillaries and arterioles, these cells are also seen in other organs (lymph nodes, spleen, heart, stomach, kidneys, and brain). Widespread fibrinoid degeneration of small blood vessels is seen in many organs, including the lungs, heart, kidneys, spleen, lymph nodes, meninges, GI tract, skeletal muscle, and bladder. Antigen specific for Hendra virus can be demonstrated by immunohistochemical staining in the vascular lesions and along alveolar walls. Intracytoplasmic viral inclusion bodies can be seen in infected endothelial cells by electron (but not light) microscopy. When respiratory disease is predominant, the principal gross lesions are severe edema and congestion of the lungs and marked dilatation of the subpleural lymphatics. The airways are filled with thick froth, which is often blood-tinged. Additional lesions seen in some affected horses include increased pleural and pericardial fluids, congestion of lymph nodes, hemorrhages in various organs, and slight jaundice.
Microscopically, the primary lesions are those of an acute interstitial pneumonia. Severe vascular damage, with serofibrinous alveolar edema, hemorrhage, thrombosis of capillaries, necrosis of alveolar walls, and alveolar macrophages are evident in the lungs.
If neurologic disease is predominant, lesions of nonsuppurative meningitis or meningoencephalitis, including perivascular cuffing, neuronal degeneration, and focal gliosis, have been seen.
Hendra virus infection should be considered when there is acute onset fever and rapid progression to death, but a nonfatal outcome should not preclude consideration of Hendra virus. Confirmation of the diagnosis is based on laboratory examination of appropriate specimens to detect virus, viral antigen, viral nucleic acid, or specific antibodies. The approach to specimen collection should reflect the serious zoonotic potential of Hendra virus and should incorporate appropriate measures to avoid human exposure. Minimum recommended samples include a blood sample (whole and/or EDTA) and nasal, oral , and/or rectal swabs. These can be taken from both live and dead horses. Necropsy specimens, both fresh and fixed in 10% formalin, of lung, kidney, spleen, liver, lymph nodes, and brain will increase the likelihood of reaching a conclusive diagnosis but also potentially increase the risk of human exposure. The number and type of specimens collected should follow a careful risk analysis by the veterinarian to prevent human exposure and consider many factors, including personal protective equipment available, training, and prior experience. If there are personal safety concerns, only a minimal set of samples (blood, swabs) should be collected. Submitting a combination of EDTA blood, serum, nasal, oral, and rectal swabs should be sufficient to detect Hendra virus infection in a horse highly suspected to be infected. Recommended procedures for safe handling of suspect Hendra cases are available at the Biosecurity Queensland web site (viewable here).
The virus can be isolated in a range of cell lines; Vero cells are the cell line of choice. Viral cytopathic effect, which typically develops after 3 days, is characterized by syncytia formation in infected cells. Virus isolation and other diagnostic tests involving live virus should only be attempted under biosecurity Level 4 conditions. Serologic confirmation of infection is based on testing acute and convalescent sera collected 2–4 wk apart in a virus neutralization test. Presence of the characteristic vascular lesions on histopathology is highly suggestive of the infection; specificity of the lesions can be confirmed by immunochemical labeling with Hendra virus reference antiserum.
African horse sickness can clinically mimic Hendra virus infection and should be considered in the differential diagnosis. Other causes of sudden death that must be excluded include anthrax, botulism, certain bacterial infections (eg, pasteurellosis, equine influenza, peracute equine herpesvirus 1 infection), snake bite, and plant or chemical poisoning.
Treatment, Prevention, and Control
There is no specific antiviral treatment for Hendra virus infection. A vaccine, containing a noninfectious protein component (G protein) of the virus, has been developed; it was introduced in November 2012 and is available through accredited veterinarians in Australia. Healthy horses can be vaccinated from 4 mo of age with two doses at a 21-day interval, followed by boosters every 6 mo. Studies are being conducted to determine whether the period of booster vaccination can be extended to 1 yr.
Confirmed cases should be euthanized on humane grounds and to limit risk of human exposure. In Australia, euthanasia of recovered seropositive horses is also recommended, because the currently available evidence cannot exclude the possibility of recrudescence in these animals.
Prevention focuses on minimizing contact with fruit bat body fluids/contaminants and includes simple, practical measures such as placing feed and water containers under cover and minimizing the number of bat food trees/shrubs (fruiting and/or flowering) in horse paddocks or excluding horses from the vicinity of such trees/shrubs. Control is based on euthanasia and deep burial of cases; monitoring, isolating, and restricting movement of in-contact animals; and disinfection of potentially contaminated surfaces.
Human infection with Hendra virus has a 57% case fatality rate. All human infections have occurred from handling infected horses (both live horses and dead horses at necropsy), so great care should be taken to ensure the personal safety of all people in contact with suspect or confirmed equine cases. Neither bat-to-human nor human-to-human transmission has been recorded.
Protocols to minimize risk of human exposure should be implemented on suspicion of Hendra virus infection in a horse, not on confirmation. An outline of the approach developed by Biosecurity Queensland includes the following steps to minimize risk. First, a plan should be made in advance that outlines how Hendra virus risks will be managed by the practice and individual veterinarians in that practice. This includes 1) taking precautions based on suspicion of Hendra virus and not waiting for confirmation of infection; 2) isolating sick or dead horse(s) from people and all other animals, including pets; 3) limiting human contact with in-contact horses to only essential people; 4) promoting personal hygiene (especially hand washing, showering) for in-contact staff; 5) identifying hazards and taking steps to minimize the risks associated with these (eg, if decontaminating an area, avoid generating splashes and aerosols by not using a high-pressure hose); 6) informing people who may be potentially exposed such as owners, handlers, and others (including other veterinarians and veterinary assistants) of the risk and the appropriate procedures to be followed; and 7) referring to relevant animal health and public health authorities.
Second, adequate personal protective equipment should be used: 1) all exposed skin, mucous membranes and eyes should be protected from direct contact; 2) inhalation of airborne particulates should be prevented; 3) regular hand washing and washing of exposed skin with soap should be promoted; and 4) cuts and abrasions should be covered by water-resistant occlusive dressings that are changed as necessary.
In particular, blood and other body fluids (especially respiratory and nasal secretions, saliva, and urine) and tissues from sick or dead horses should be treated as potentially infectious and appropriate precautions taken to prevent any direct contact with, splashback of, or accidental inoculation with these body fluids.
Last full review/revision October 2014 by Nina Yu-Hsin Kung, PhD, MSc, BVSc, BVM