West Nile and Other Flavivirus Encephalitis

Robert J. MacKay

The causative agent of West Nile encephalitis (WNE) is a small, positive-sense, single-stranded, enveloped RNA virus,¹ West Nile virus (WNV), which is a member of the Japanese B encephalitis antigenic complex of neuroinvasive arboviruses within the genus Flavivirus, family Togaviridae.¹ In addition to WNE virus, Kunjin, Japanese B encephalitis, and Murray Valley encephalitis viruses are members of the Japanese B encephalitis antigenic complex that have caused encephalo­myelitis in horses outside the United States.

The Kunjin virus is a lineage 1 WNE virus that traditionally has caused mostly subclinical or febrile nonneurologic disease in horses and humans in northwestern Australia and Papua New Guinea. A neurovirulent strain of Kunjin virus emerged in New South Wales in Australia in the late summer and fall of 2011 and caused encephalitis in at least 1000 horses in southeastern Australian states, with mortality rates of 10% to 15%.^4,5 This unprecedented epizootic of Kunjin encephalitis (and also of encephalitis due to the arboviruses Ross River virus and Murray Valley encephalitis virus) followed unusually heavy spring rainfalls in southeastern Australia.⁴ There is evidence that the 2011 outbreak probably resulted from a complex interplay among novel genetic virulence determinants and unusual ecologic and epidemiologic conditions.⁵ Since the Murray Valley encephalitis epizootic in 1974, cases had been reported only sporadically.⁶ Japanese B encephalitis virus is endemic in east Asia and southeast Asia and has made incursions into the northern tip of Australia.⁷ It is maintained in cycles among porcine and avian reservoirs through mosquito hosts.

■ HistoryandDistribution WNV was originally isolated from the blood of a febrile woman in the West Nile region of Uganda in 1937. Cases of WNE in humans were first recorded in Israel in the 1950s, and cases in horses were recorded in Egypt and France between 1960 and 1965.^8,9 Since 1996, cases have been sporadic, and outbreaks of equine WNE have been small and localized in numerous countries in the Mediterranean basin and northern Europe. These incursions of lineage 1 and, more recently, lineage 2 WNV continue to expand in number and geographic range.¹⁰ It is now clear that WNV has become enzootic in several locations in Europe.

In August 1999, the first case of WNE in humans was reported in Queens, New York. Characterization of the virus from this outbreak showed it was highly homologous with Israel 1998 (Is98), a strain of WNV that had originally been isolated from a goose.¹¹ Ultimately, 62 cases occurred in humans in 1999, with 7 mortalities. From August to October 1999, 25 horses on Long Island were diagnosed with WNE; 9 of these horses died.¹² Deaths occurred in numerous native bird species, particularly in corvids (crows, ravens, jays) and in exotic species housed at the Bronx zoo. In 2001 the disease became a national epizootic; cases occurred for the first time in Florida, Kentucky, Louisiana, Georgia, Alabama, Illinois, Indiana, Mississippi, New Hampshire, North Carolina, Tennessee, and Virginia.¹³ There were 738 cases reported nationally, of which 33.2% died or were euthanized.

Equine or human WNE or both have been reported in multiple Canadian provinces, and there is serologic evidence of WNV in the Caribbean and Latin America.¹

■ Natural History WNV cycles between ornithophilic mos­quito vectors and avian reservoir hosts. In temperate climates, the virus overwinters by long-term infection of hibernating vector mosquitoes, followed by vertical (transovarial) transmis­sion and possibly by latent infection of avian hosts.¹⁴ Culex spp. are the most important vectors for the North American epizootic in birds. Different species transmit the virus in different parts of the country. Major vector species include Culex pipiens in the northeast and central United States, C. tarsalis from the far western United States to the panhandle of Texas, and Culex quinquefasciatus in the southeastern United States from Texas to Florida. Although birds of more than 300 species have been reported to be WNV-positive, passerine (true perching bird) and charadriiform (shorebird) species are considered the most competent reservoir hosts for WNV in the United States. Corvid species such as crows and blue jays are efficient hosts, but their mortality rates were high during the North American epizootic. The house sparrow (an Old World passerine species) also develops high-titer viremia after WNV infection but is relatively resistant to the adverse effects of the virus.

“Bridge vector” species of mosquito transmit the virus from birds to mammals, including horses. The mosquito species important for bird-horse transmission are not known with certainty but may include members of C. pipiens complex, C. tarsalis, and Culex salinarus.

The occurrence of WNE reflects vector mosquito activity, which is seasonal in temperate zones and year-round in tropical and subtropical regions. Significant viral activity begins in July, and the incidence peaks in September and October. The epizootic usually declines rapidly after the first frost.

■ Clinical Signs Experimental infections in horses result in viremia 3 to 5 days after viral inoculation into the CSF, and clinical signs begin at 7 to 10 days.¹⁴ Experience during clinical outbreaks indicates a similar 1- to 2-week incubation period before signs of WNE are manifested.

Clinical signs have been extremely varied; however, surveys of cases published across the United States and from Europe have revealed some consistent combinations of signs.^8,12,16-20 Approximately 25% of affected horses have had mild to moderate fever at the time of initial examination, although it is likely that most were febrile before examination. Consensus neurologic signs in 726 cases from the northeastern United States, Florida, Indiana, Nebraska, and Colorado^8,12,16-20 have included limb ataxia (71%), limb weakness (58%), muzzle twitching (45%), obtundation (43%), recumbency (31%), lip droop (22%), hyperresponsiveness/hyperesthesia (18%), odontoprisis (12%), dog-sitting posture (10%), thoracic limb collapse (7%), compulsive walking (7%), muscle atrophy (6%), seizure (4%), blindness (4%), circling (4%), and head pressing (3%).

The observation of coarse twitching of the muzzle (and, to a lesser extent, the eyelids) is a particularly distinctive finding in acute cases of WNE. Less frequent are fine fasciculations of the cutaneous muscles of the torso and neck. Horses acutely affected may be frantic, fearful, and hyperresponsive to stimuli. This stage is transient (usually 6 cells/uL) without significant red blood cell (RBC) contamination, neutrophil proportions ranged from 0% to 8% (median, 2.5%; H. L. Wamsley, personal communica­tion, 2007). Of interest is that nucleated cell counts and protein concentrations were significantly higher in lumbosacral samples than in atlantooccipital samples.²³ The CSF findings have not been strongly correlated with the severity of the clinical presentation.

■ Diagnosis A presumptive diagnosis is often based on clinical signs, especially in areas where the disease or virus is currently active and the horse has not been vaccinated against WNV within the previous 12 months. Diagnostic confidence is increased by the observation of muzzle twitching. The finding in a CSF sample of mild to moderate mononuclear pleocytosis with low numbers of neutrophils further supports a diagnosis of WNE. In a horse with signs of encephalomyelitis, presump­tive diagnosis requires a positive result of a serum IgM antibody capture-ELISA and negative or unsupportive results of tests for other possible diagnoses, including alphaviral encephalitis, equine herpesvirus myeloencephalopathy, EPM, and hepato- encephalopathy. Single-point assays of anti-WNV IgG, such as the plaque-reduction neutralization test, are not particularly useful because titers can be affected by prior vaccination; however, demonstration of a fourfold or greater rise in titer between acute and convalescent samples supports the diagnosis in unvaccinated horses.

At necropsy, PCR (nested or quantitative), immunohisto­chemical, and viral isolation methods are used to demonstrate the presence of viral components in the CNS,¹⁴ although immunohistochemistry tests are much less sensitive for WNV than for EEE.

Differential diagnoses for WNE in horses include alphaviral encephalomyelitis (EEE, WEE), rabies, verminous or bacterial meningoencephalomyelitis, EPM, brain or spinal cord trauma, cervical stenotic vertebral myelopathy, hepatic encephalopathy, or compressive mass in the calvaria or vertebral canal.

■ Treatment No antiviral treatment for horses with WNE is readily available. Type I interferons (i.e., interferon α or β) or interferon inducers stimulate nonselective cellular resistance to viral infection and are used for prevention or treatment of viral diseases. The limited data so far reported for the use of interferon in humans with WNE do not suggest any positive effect on outcome, although some anecdotal evidence supports this approach in horses. A reasonable dosage is 3 ? 106 to 15 ? 10⁶ international units (IU) of interferon β_1α IM or IV (Avonex [Biogen, Cambridge, Mass.], 30 μg) three times weekly. The concept of therapy with WNV-specific globulin is supported by findings in a guinea pig model of WNE, although no controlled studies of its use in horses have been reported. In a limited series of cases, horses given a hyperimmune globulin product showed a trend for less recumbency, but the product had no significant effect on severity of clinical signs or mortal­ity.¹⁴ Several commercial products of this type are available for horses in the United States. These products are given slowly IV according to label instructions.

Supportive care is of paramount importance. Secure footing, deep dry bedding, protective covering of distal limbs and bony prominences, climate control, and provision of ample food and water are important in all cases. Recumbent horses must be repositioned from one side to the other regularly, and decubital injuries must be addressed. Early and systematic use of slinging is helpful in some otherwise recumbent horses. The Anderson sling, which transfers some of the horse’s weight onto the upper limbs, is probably more effective in horses with WNE than are slings that rely entirely on torso support. In one series of 12 horses with prolonged recumbency that were treated by slinging, 2 were able to stand unassisted after slinging was used to help them rise, and 2 were able to stand after slinging was discontinued.¹⁸ In horses that have long-term difficulty in rising to the standing position, semipermanent mobile slinging systems have been devised and marketed. These slings are stocklike frames that have wheels on each corner and webbing that rides under and supports the horse's torso.

NSAIDS and other antiinflammatory drugs are typically used empirically in horses with WNE to control local pain, inflammation, and fever. Drugs commonly used for this purpose are flunixin, 1.1 mg/kg IV or PO (bid in U.S., sid in Europe); DMSO, 0.25 to 1 g/kg in a 10% solution IV or intragastrically bid; vitamin E, 20 IU/kg PO q24h; and pentoxifylline, 10 mg/ kg PO bid. The use of corticosteroids is controversial. Because recrudescence of WNV infection and clinical relapse in recover­ing immunosuppressed horses are at least theoretical concerns, corticosteroids should be used only in horses that are recumbent or in imminent danger of becoming so. A typical dosage regimen is dexamethasone, 0.05 to 0.1 mg/kg IV or 0.1 to 0.2 mg/kg PO, once or twice daily for 1 to 3 days. Affected horses are often treated for presumptive EPM with antiprotozoal drugs (e.g., ponazuril, 5 mg/kg PO q24h) until IgM antibody capture-ELISA results become available.

Recovery, when it occurs, takes from 1 day to several months (mean, 22.3 days for 271 horses recovered from WNE in Nebraska and Colorado¹⁸ and 10.7 days for 569 horses in North Dakota²⁰) and is not always complete. Of 125 predominantly pleasure horses in Minnesota that survived WNE, 22 (17.6%) were classified by their owners as incompletely recovered 6 months after diagnosis.²⁴ Gait abnormalities, behavioral changes, muscle atrophy, and lethargy were among the residual abnor­malities reported by the horses' owners. An additional 29 horses (23.2%) were considered fully recovered with residual deficits. A surprisingly high 9% of horses suffered some form of relapse of clinical signs 2 weeks to 5 months after diagnosis.

■ Necropsy Findings Gross abnormalities of the CNS in horses with WNE are limited to meningeal congestion and hemorrhagic foci in the brain and spinal cord. Histologic study reveals mild to severe multifocal lymphocytic polioencepha­lomyelitis with a predilection for the ventral and lateral horns of the thoracolumbar spinal cord. The brain exhibits lymphocytic perivascular cuffing and microgliosis in the medulla oblongata and pons and, to a lesser extent, in the basal nuclei, thalamus, and mesencephalon. In severe cases, there is evidence of neuronal degeneration characterized by central chromatolysis, cytoplasmic swelling, or cell shrinkage. The cerebral and cerebellar cortices usually are relatively free of lesions.

■ Control General mosquito control is important at the individual operation level. Local authorities may undertake area-wide mosquito abatement when sentinel chicken serologic tests or viral isolations from wild birds suggest high local WNV activity.

A killed whole virus vaccine has been available since August 1, 2001, and has had full U.S. Food and Drug Administration approval since 2003. This vaccine is effective beginning approximately 2 weeks after the second dose of the initial series, but antibody levels drop substantially by 4 months after vaccination. Two additional vaccine types became available afterwards: a recombinant subunit vaccine in a pigeon poxvirus vector and a live chimera virus vaccine based on an attenuated yellow fever virus backbone. WNV vaccines are available in combinations with alphaviral encephalitis and tetanus vaccines, although there is evidence that simultaneous multiple vaccines can diminish the antibody response to WNV vaccination.²⁵ The live chimera vaccine was withdrawn in 2010 for safety reasons, and an inactivated version is now offered. For all products, the primary series or booster vaccination is completed at least 1 month before anticipated WNV activity, and boosters are given once or more annually, depending on the length of the vector season. Various challenge models have been used for each of these vaccines (except the killed version of the chimera vaccine) to demonstrate early protection against encephalitis and continuing protection against infection 12 months after a primary vaccination series. Foals of vaccinated mares should begin vaccination at 4 to 6 months of age and should have been given three inoculations by 12 months of age. Boosters should be given at least annually and more often in high-risk areas. Vaccine failures are more common in old horses. A DNA vaccine against WNV was approved by the U.S. Department of Agriculture in July 2005 but has since been discontinued.