Chapter 1 West Nile Virus in Birds and Mammals
West Nile virus (WNV) is an arthropod-borne virus (arbovirus) in the family Flaviviridae, genus Flavivirus—Japanese encephalitis antigenic complex—that includes Alfuy, Cacipacore, Japanese encephalitis, Koutango, Kunjin, Murray Valley encephalitis, St. Louis encephalitis, Rocio, Stratford, Usutu, West Nile, and Yaounde viruses. Flaviviruses share a common size (40-60 nm), symmetry (enveloped, icosahedral nucleocapsid), nucleic acid (positive-sense, single-stranded RNA of ∼10,000-11,000 bases), and appearance on electron microscopy. The close antigenic relationship of the flaviviruses, particularly those belonging to the Japanese encephalitis complex, accounts for the cross-reactions observed in diagnostic serologic assays.31
WNV has been described in Africa, Europe, the Middle East, West and Central Asia, Oceania (subtype Kunjin), and most recently in the Western Hemisphere. It was first isolated from a febrile adult woman in the West Nile District of Uganda in 1937, and its ecology was first characterized in Egypt in the 1950s. The virus became recognized as a cause of severe human meningitis or encephalitis in Israel in 1957. Equine disease was first noted in Egypt and France in the early 1960s.33,40,57 Recent outbreaks of WNV encephalitis in humans have occurred in Algeria in 1994, Romania in 1996-1997, the Czech Republic in 1997, the Democratic Republic of the Congo in 1998, Russia in 1999, the United States (U.S.) in 1999-2003,7,10,19,58 and Israel in 2000.33 Epizootics have occurred in horses around the Mediterranean (Morocco in 1996, Italy in 1998, France in 2000),57 and in the U.S. in 1999-2001.61,78 A thorough review of pre–North American WNV ecologic history was published by Komar.40
WNV was first found in North America in New York City in humans, equines, and free-ranging and captive wildlife in 1999.7,40,58,71 Since 1999 in the U.S., more than 20,000 humans have been infected, causing more than 700 deaths, and more than 23,000 equine cases and hundreds of thousands of avian cases have been reported. Most cases occur in North America during the summer and fall between July and October, with peaks in August and September. Spread across North America to all 48 contiguous states and seven Canadian provinces has been documented in twice-weekly summary reports available on the U.S. Centers for Disease Control and Prevention (CDC) World Wide Web site* and by interactive maps collated by the U.S. Geologic Survey.† In Canada, Health Canada summarizes WNV activity.‡ Since 1999, surveillance data have shown WNV activity in the Cayman Islands in 2001 (CDC); birds in Jamaica23 and Guadeloupe64 in 2002; horses, humans, and wildlife in Mexico in 20026,24,49; and birds in the Dominican Republic in 200341 and in Puerto Rico and El Salvador in 2004.17
The arboviral encephalitides are zoonotic, being maintained in complex life cycles involving a nonhuman primary vertebrate host and a primary arthropod vector. Transmission occurs between susceptible vertebrate hosts by blood-feeding arthropod mosquitoes, sand flies, ceratopogonids, “no-see-ums,” and ticks.1,2,48,68 Infection usually occurs as a result of a mosquito bite while taking a blood meal. Normal transmission cycles usually remain undetected until humans or other mammals become “accidentally” infected, potentially as the result of some ecologic change. Humans and domestic animals may develop clinical illness but usually are incidental or “dead-end” hosts because they do not produce significant viremia and thus do not contribute significantly to the transmission cycle.
Since 1999, more than 60 separate species of mosquito have been positive (virus isolated, RNA or antigen detected) through national surveillance.* Although not all these are competent vectors, the predominant species testing positive are Culex spp. The discovery that hybrid Culex mosquitoes sometimes feed on both humans and birds resulted in a focus on potential “bridge vectors.”27 One risk assessment of mosquito feeding characteristics identified Culex pipiens and C. restuans as the most competent vectors for humans.37 A 5-year analysis of mosquito data in Connecticut revealed that Culex spp. were the most prevalent carriers from July to September, playing a roll in early-season enzootic transmission and late-season epizootic amplification in wild birds. Culex restuans was most prevalent in June and July and may play an important role in enzootic transmission and amplification in wild birds early in the season. Culiseta melanura was found to be the major orniphilic species and may play a major role in amplification among birds. Aedes vexans may play a significant role in transmission to mammals.2
Other, non–arthropod-borne routes of transmission have been reported. New transmission routes in humans include infection through contaminated blood products and transfusion13,63 and organ transplantation,14 maternal transmission through breast milk and intrauterine transmission,12 and occupational exposure through laboratory “sharps.”11 Experimentally, infection has been demonstrated after oral exposure in cats fed infected mice and birds.3,42 Oral exposure to horse meat is the hypothesized route of transmission for infected alligators.55 Fecal shedding was identified as a potential route of transmission after experimental direct transmission between cage mates in crows (Corvus brachyrhynchos), blue jays (Cyanocitta cristata), black-billed magpies (Pica pica), and ring-billed gulls (Larus delawarensis).42,54 Experimental and natural direct transmission also occurred between geese.4,75 The importance of these transmission routes is unknown but thought to be of secondary significance compared with arthropod-borne transmission for amplification and spread of the disease.
The WNV strain first identified in New York was closely related to that recently isolated in Israel.33,46 Although the route of introduction is not known, hypotheses include release of infected vectors or hosts through international commerce or travel. Geographic spread via introduction through migratory birds has been hypothesized51,65,66 worldwide but is unlikely in the case of introduction into North America. A quantitative risk assessment of pathways by which WNV could reach Hawaii suggests that the most viable hosts for introduction are mosquitoes, rather than birds or other hosts. Viable routes of introduction include transfer of infected hosts via plane or boat, and introduction through migratory birds could not be quantified in this case.36
Cases are confirmed by combining clinical and laboratory criteria. Standard clinical and laboratory case definitions have been derived for humans7,9,30,52,63 and are updated periodically on the CDC website.*
Arboviral infections may be asymptomatic or may result in febrile illnesses of variable severity sometimes associated with central nervous system (CNS) involvement (aseptic meningitis, myelitis, and encephalitis). Arboviral meningitis is usually characterized by fever, headache, stiff neck, and pleocytosis in cerebrospinal fluid (CSF). Arboviral myelitis is usually characterized by fever and acute paresis or flaccid paralysis. Arboviral encephalitis is usually characterized by fever, headache, and altered mental status ranging from confusion to coma with or without additional signs of brain dysfunction. Nonneuroinvasive syndromes include myocarditis, pancreatitis, or hepatitis. In addition, arboviral infections may cause febrile illnesses (“West Nile fever”) with headache, myalgias, arthralgias, and sometimes accompanied by skin rash or lymphadenopathy.
Because closely related arboviruses exhibit serologic cross-reactivity, positive results of serologic tests using antigens from a single arbovirus may be misleading, as in determining that antibodies detected against St. Louis encephalitis virus are not the result of an infection with West Nile (or dengue) virus, or vice versa, in areas where both these viruses occur. In areas where WNV has circulated in the recent past, the coexistence of WNV-specific IgM antibody and illness in a given case may be coincidental and unrelated. In those areas the testing of serially collected serum specimens assumes added importance.
Most cases in animals are defined through isolation of the virus or detection of genetic material postmortem. Virus isolation is the “gold standard” but takes time, which limits its use as a rapid surveillance tool. Immunohistochemistry (IHC) and reverse-transcription, nested polymerase chain reaction (RT-PCR) tests are used for detection of antigen, and an antigen-capture assay is also commercially available. No definitive list exists for the most effective combination of tissue and test methodology by species, but some optimal combinations have been reported.82,83 In general, virus is best detected in kidneys, brains, and hearts, as well as on oropharyngeal or cloacal swabs antemortem. The success of IHC depends greatly on tissue selection (heart, kidney, liver, lung); brain tissue is best for virus isolation; and RT-PCR is generally the most sensitive test for all tissues, with few reported exceptions.* Recently, feather pulp collected from bird carcasses has been shown to be useful in dead-bird surveillance when tested by RT-PCR.22
Standard antibody detection methods include the hemaglutination inhibition (HAI) test and plaque reduction neutralization test (PRNT). The HAI test is hindered by nonspecific reactivity, whereas the PRNT is more specific and may differentiate between antibody reactivity from closely related viruses.30 In the U.S. the PRNT requires biosafety level 3 facilities, which may limit its use in many laboratories. IgM-capture enzyme-linked immunosorbent assays (ELISAs) have been developed for humans, equines, canines, and chickens and are useful for determining recent exposure to the virus.30 The IgM ELISA may be used as a screening test, with PRNT performed to differentiate between St. Louis encephalitis and WNV as confirmation. A blocking ELISA was developed for broad species use and is used to determine the origin of antibody reactivity. A broadly reactive IgG-capture ELISA for bird serum has proved to be effective in a wide variety of birds but needs to be confirmed by PRNT.29,30 Antibody persistence in naturally and experimentally exposed birds is variable. In a study of wild-caught rock pigeons (Columba livia) naturally infected with WNV, antibodies were found to persist for longer than 15 months, as detected by ELISA and PRNT; maternal antibodies persisted for an average of 27 days. Both tests outperformed the HAI test.29
WNV has an extremely broad host range, replicating in birds, reptiles, amphibians, mammals, mosquitoes, and ticks.30 Reviews of pathologic findings in various animal species are available.*
Cases of WNV disease in horses have been documented either by virus isolation or by detection of WNV-neutralizing antibodies every year since 1999. WNV infection in horses and other domestic equids ranges from asymptomatic to fatal encephalitis. Common clinical signs include ataxia, incoordination, lethargy, weakness, hind limb paresis, muscle tremors and fasciculations, recumbency, and death. Experimental studies suggest that about 10% of infected horses develop clinical illness.41,61 From 20% to 40% of equine WNV cases result in death or euthanasia.69,70,79,80 Horses most likely become infected by the bite of infectious mosquitoes. In a review of 569 cases, the risk of death among nonvaccinated horses was 3 to 16 times higher than in vaccinated horses after one or two doses.70