Detection of WN virus in wild bird populations is a useful indicator for predicting and preventing infection in humans and domestic animals (horses). (D67, J115.13.w2)

Timing and level of surveillance:

• Northeastern and Midwestern USA: Active ecological surveillance should begin in early spring and continue through the fall (autumn) until mosquito activity ceases because of cold weather. (D147)
  • During 2001-2002, onset of human illness in the northeastern states occurred from early July to mid-November; in the same years, cases in birds occurred as early as the first week of April and as late as the second week of December. (D147)
• Southern USA: Active ecological surveillance should be conducted year round in these areas. (D67, D147)
  • N.B. in the Southern states, WN virus circulates throughout the year, with activity detected from January to December. In 2001-2004, onset of human illness was reported from as early as April to as late as mid-December; equine and avian infections were reported in all months of the year in these geographical areas. (D147, P39.3.w1, J84.10.w5, J84.11.w4)
• Western USA:  Ecological surveillance should be encouraged beginning in early spring and continuing through the fall until mosquito activity ceases due to cold weather. (D147)
  • In 2002, WNV activity was first reported among humans and animals in Rocky Mountain states and among animals in Pacific coast states. These events occurred relatively late in the year (mid-August). Predicting the temporal characteristics of future WNV transmission seasons in these geographical regions based on these limited reports is not possible. (D147)
• Western Hemisphere outside the USA: Surveillance in other countries of the Western Hemisphere should be encouraged. (D67)
  • In Canada, in 2002 a WNV epidemic occurred in Ontario and Quebec provinces and an equine/avian epizootic occurred extending from the maritime provinces to Saskatchewan. By 2005, WV virus had been detected in Alberta, Manitoba, New Brunswick, Nova Scotia, Ontario, Quebec and Saskatchewan. (B526.21.w21, D147)
  • Development of surveillance systems capable of detecting WNV activity should be encouraged in the Caribbean and Central and South America. WNV surveillance should be integrated with dengue surveillance in these areas, and with yellow fever surveillance in areas where urban or peri-urban transmission of this virus occurs. (D147)
    • Wild bird and sentinel chicken surveillance has been recommended as part of surveillance for WNV in the Caribbean and in South America (Brazil). (J408.40.w1, W705.May08.w1)
    • By 2005, WV virus had been detected in equines and/or birds in Mexico, the Caribbean (the Bahamas, the Cayman Islands, Cuba, the Dominican Republic, Guadeloupe, Jamaica, Puerto Rico, Trinidad), Central America (Belize, Guatemala, El Salvador), and South America (Colombia); in 2006 it was detected in Venezuela and caused fatal illness in equines in Argentina; samples taken from birds in Argentina were first positive in 2005. (B526.21.w21, J84.12.w3, J84.13.w1, J84.14.w1, J270.45.w1, J491.19.w1)
• Southern Africa: transmission of WNV increases in the early months of the year, following heavy rainfall in spring and summer. (J84.11.w4)

Sample collection, labelling and shipping: 

It is important to ensure that the correct protocols are followed for sample collection, labelling and shipping of specimens. Current information regarding these should be checked, generally with the receiving laboratory, before collection and packaging of samples. For example, information regarding labelling and shipping containers for samples to be sent to CDC for WN virus testing may be found in CDC Epidemic/Epizootic West Nile Virus in the United States: Guidelines for Surveillance, Prevention and Control 3rd Revision (D147 - Appendix A), while information regarding shipping of specimens to the New York State Department of Environmental Conservation's Wildlife Pathology Unit is contained in the New York State West Nile Virus Response Plan - Guidance Document (D72) Appendix E: Guidelines for Reporting and Submitting of Birds and Mammals, West Nile Virus, 2001. 

• CDC requests that avian brain, heart, kidney and spleen samples be submitted fresh-frozen. (D147 - Appendix A)
• Details of requirements for samples sent to CDC are provided in: West Nile Virus - Detection and Identification Techniques (Viral Reports) - Sample Collection & Shipping

Laboratory case definitions

The following case definitions for laboratory-confirmed and laboratory-probable WN virus infection are taken from CDC Epidemic/Epizootic West Nile Virus in the United States: Guidelines for Surveillance, Prevention and Control 3rd Revision (D147): 

Sentinel species:

• Laboratory-confirmed WNV infection: 
  • WNV isolation, RNA detection, or antigen detection:
    • WNV isolation (identity of virus established by at least two of the following techniques:
    • Positive RT-PCR test for WN viral RNA with validation by 1) repeated positive test using different primers, 2) positive PCR result using another system (e.g., TaqMan), or 3) virus isolation.
    • Detection of WN viral antigen (e.g., IFA, EIA, VecTest^TM) validated by inhibition test (for ELISA), RT-PCR, or virus isolation
  • Seroconversion to WNV in serially collected serum specimens, by plaque-reduction neutralization**
  • Detection of IgM antibody to WNV, validated by demonstration of neutralizing antibody to WNV**
    • ** SLE virus infection should be ruled-out by cross-neutralization; criterion for PRNT positive is a 90% neutralization titer of at least 1:10, and 4-fold greater titer compared to other flaviviruses such as SLE.
• Laboratory-probable WNV infection:
  • Detection of IgM antibody to WNV; or
  • Seroconversion to WNV in serially collected serum specimens, strongly reactive by EIA or IFA.

Avian mortality (dead bird): 

• Laboratory-confirmed WNV infection
  • WNV isolation, RNA detection, or antigen detection:
    • WNV isolation (identity of virus established by at least two of the following techniques:
    • Positive RT-PCR test for WN viral RNA with validation by 1) repeated positive test using different primers, 2) positive PCR result using another system (e.g., TaqMan), or 3) virus isolation.
    • Detection of WN viral antigen (e.g., IFA, EIA, VecTestTM) validated by inhibition test (for ELISA), RT-PCR, or virus isolation
• Laboratory-probable WNV infection:
  • Positive RT-PCR test for WN viral RNA in a single test; or
  • Antigen detection not validated by another procedure

(D147)

Record keeping:

Official data recording sheets should be used in order to standardise information and facilitate both entry of information into databases and comparison of data from different areas or collected by different personnel. An example of a data recording sheets for reporting dead/ill birds is given in the New York State West Nile Virus Response Plan - Guidance Document (D72) Appendix D: Surveillance Report Forms.

• The CDC Epidemic/Epizootic West Nile Virus in the United States: Revised Guidelines for Surveillance, Prevention and Control (D67) emphasizes that accurate taxonomic identification of specimens, a unique identification numbering system for specimens and durable tagging of filed specimens, as well as standardized forms for data collection and specimen submission, are particularly important when dealing with wild birds (and mosquitoes). (D67)

NOTE: 

• USA, 1999-2007. WN virus was detected in 1999 in 249 wild birds; in 2000 in 4,305 wild birds; in 2001 in 7,332 wild birds; in 2002 in 16,741 wild birds; in 2003 in 12,066 wild birds; in 2004 in 7,396 wild birds; in 2005 in 5,393 wild birds; in 2006 in 4,106 wild birds; and in 2007 in 2,182 wild birds. (V.w118 - data from CDC-ArboNET)
  • Data on WN virus in wild birds in the USA to 2007 is available summarised in map form. See: Map0001 - Spread of West Nile Virus in the USA (2000-2007) - Bird maps
• USA, 2000-2007. WN virus was detected in 2000 in 13 sentinels; in 2001 in 218 sentinels; in 2002 in 1,640 sentinels; in 2003 in 1,956 sentinels; in 2004 in 1,676 sentinels; in 2005 in 1,651 sentinels; in 2006 in 917 sentinels; and in 2007 in 904 sentinels. (V.w118 - data from CDC-ArboNET)
  • Data on surveillance using sentinels in the USA to 2007 is available summarised in map form. See: Map0001 - Spread of West Nile Virus in the USA (2000-2007) - Sentinels maps

• Guidelines for Arbovirus Surveillance Programs in the United States

• Epidemic/Epizootic West Nile Virus in the United States: Revised Guidelines for Surveillance, Prevention, and Control

• Hazard Assessment of the Organophosphates

• New York State West Nile Virus Response Plan - Guidance Document

• Health Canada - Municipal Mosquito Control Guidelines

1) the timely reporting and analysis of dead bird sightings; and 

2) submission of selected individual birds for WN virus testing. 

(D147)

The goal of dead/sick wild bird surveillance is to utilize bird mortality associated with WN virus infection as a means of detecting WN virus activity in a given location. (D67)

Protocols and specimens:

Video Available: Bird Necropsy Protocol for West Nile Virus Surveillance:
Internet (Web) Version (Smaller files - quicker to load)
CD-ROM Version (Larger files - higher quality images)

SPECIES IDENTIFICATION: The image below provides a comparison of black corvids (Corvidae: the crow family) likely to be submitted in the USA for WNV surveillance

The level of effort of wild bird surveillance will depend on the risk assessment for each jurisdiction:

• In general, avian surveillance should be initiated when local adult mosquito activity begins in the spring;
• A database should be established to record and analyze dead bird sightings using the following suggested data:
  • caller identification and call-back number, 
  • date observed, 
  • location geocoded to the highest feasible resolution, 
  • species,
  • condition. 
• Birds in good condition (without obvious decomposition, scavenging or infestation with maggots) may be submitted for laboratory testing;
  • Carcasses should be handled carefully, avoiding direct contact with skin. 
• For greatest sensitivity, a variety of bird species should be tested, but corvids (Corvidae) should be emphasized;
• The number of bird specimens tested will be dependent upon resources and whether WN virus-infected birds have been found in the area;
• Triage of specimens may be necessary on the basis of sensitive species and geographic location.
• A single organ specimen from each bird is sufficient to detect WN virus or viral RNA. 
  • Kidney, brain or heart is preferable. 
  • Kidney has been the most consistently positive tissue at all levels of infection in crows (Corvus brachyrhynchos - American Crow), with brain the next most consistent organ. (J133.951.w12)
    • For the TaqMan RT-PCR the kidney appears to be the optimal tissue in corvids and processing of kidney is easier than processing of heart. (J5.47.w2)
  • In owls (Strigiformes - Owls, Frogmouths & Nightjars (Order)) heart and kidney appear to be the optimal tissues for immunohistochemical staining. (J5.47.w2)
  • Heart, liver and kidney may be more suitable for testing than brain; only 40% of brain samples shown to contain virus were positive by immunohistochemistry (IHC) while overall IHC agrees with virus isolation findings in 94% of cases. (P48.4.w13)
  • N. B. Research published in May 2004 suggests that, when available i.e. during the moult, samples of feather pulp from wing or tail flight feathers may be suitable for virus detection as this material contains relatively high titres of virus, at least in corvids (Corvidae - Crows, Birds-of-Paradise etc. (Family)) and was found to be more sensitive for virus detection than either spleen/kidney samples or cloacal swabs. (J84.10.w2)
• Testing involves isolation of infectious virus, or specific RNA detection by RT-PCR, and will generally identify an infection within one to two weeks after transmission. 
• For confirmation of initial positive findings in a new geographic area, additional testing is encouraged.
• It is suggested that, because surveillance based on dead wild birds is the most sensitive method for detection of WN virus activity in most regions, surveillance of avian mortality should continue in each region "for as long as it remains necessary to know whether local transmission persists." (D147)

(D67, D147, J5.47.w2, J133.951.w12, P48.4.w13)

Advantages of avian morbidity/mortality surveillance:

• Certain species of birds, in particular corvids (Corvidae), e.g. crows and jays, appear to experience high clinical attack rates.
• The size and coloration of certain dead birds make them conspicuous (e.g., crows Corvus spp.).
• RT-PCR can be used to rapidly detect WN viral RNA in bird tissues, even if they are partially decomposed.
  • RT-PCR and antigen-detection assays have been adapted for field applications. (D147)
• Due to public involvement in reporting dead bird sightings, dead wild birds are readily available over a much wider region than can be sampled by other surveillance methods.
• Detection of WN virus-positive dead birds is likely to indicate that local transmission of the virus is occurring. 
• Provides temporally and spatially sensitive detection of WN virus activity.
• Multiple findings of WN virus in dead birds likely represents local transmission.
  • Data showing infection in mosquitoes and at least one hatch-year bird from sites in New York at which a single WN virus-positive dead bird had been found early in the previous transmission season suggest that the presence of even a single virus-positive dead bird may be evidence of local transmission of the virus and is unlikely to be due to a bird dispersing a considerable distance from the location at which it was infected. (J91.67.w1)
• Can be used for early detection and possibly also for ongoing monitoring of WN virus transmission.
• May be used to estimate risk of human infection.
• Avian mortality surveillance is important particularly for early detection of WN virus transmission in new geographical regions. (P39.3.w7)
• Dead bird density, plotted geographically, may be useful for directing efforts for mosquito (larval and adult) surveillance and control. (P39.3.w12)
• The presence of dead bird clusters may act as an early warning system for WN virus activity. (J84.9.w6)
• Finding of WNV-positive birds early in the transmission season may act as an indication that human infections are likely in the same area later. (J84.9.w14)
• "A rapid WNV reporting system using GIS/GPS data entry and VecTest of oral/cloacal swabs could provide excellent surveillance data." (P39.4.w5)

(D67, D72, D147, J84.7.w5, J84.9.w6, J84.9.w14, J91.67.w1, J133.951.w8, P39.3.w7, P39.4.w5)

Disadvantages of avian morbidity/mortality surveillance:

• It is difficult to compare dead bird surveillance data between jurisdictions.
• Quantification of infection rates is infeasible.
  • Quantification of infection rates in a given species e.g. Corvus brachyrhynchos - American Crow  would require data on the size of the local population of that species and the mortality rate following infection. (J133.951.w10)
• Birds are highly mobile and often have extensive home ranges; therefore the site of death may be distant from the site of infection. This is particularly the case outside the breeding season, when many species are less territorial. (D147)
• Collection, handling, shipping, and processing of birds or their clinical specimens are cumbersome.
• Systems for handling, processing, and testing have at times been overwhelmed by high public response and public expectations.
• The long-term usefulness of this system is uncertain because:
  • natural selection for disease-resistant birds may occur;
  • populations of susceptible species may become very low; or 
  • the virus may evolve, resulting in low or no avian mortality.
  • It may be less useful in areas where the virus recurs annually. (D147)
• Success is influenced by public participation, which is highly variable, and depends on factors such as the amount of public outreach programs and public concern.
  • In the future there may be a loss of public interest in reporting dead birds.
• The system may be less sensitive in rural areas, where there are fewer people present to observe dead birds over a wider geographic area.
  • In the western USA there is not only lower observer density but also a vector is present (Culex tarsalis) which is less ornithophilic, so that there are fewer reports of dead birds relative to other surveillance indicators not involving birds. (D147)
• In areas where other closely related flaviviruses (such as St Louis encephalitis virus) occur, birds may have cross-reactive immunity due to St. Louis virus infection, or possible a greater innate resistance to severe flavivirus infection. (J133.951.w10)

(D67, D72, D147, J133.951.w8, J133.951.w10)

Recent experience:

• Analysis of avian morbidity and mortality data 2000-2002 indicated that Corvus brachyrhynchos - American Crow was the most sensitive species for avian morbidity/mortality surveillance in northern regions of the USA. However, some areas did not have WN virus-positive American crows ((Corvus brachyrhynchos - American Crow)), but only positive birds of other species. In southern regions Cyanocitta cristata - Blue jay have been more sensitive than crows for avian morbidity/mortality surveillance. (D147)
• Almost all of the WN virus-positive birds were found singly and not as part of a mass die-off in a single time and place.
• Approximately a third of the positive birds had signs of trauma on necropsy.
• Many positive birds did not have pathology indicative of WNV infection on necropsy. No lesions are pathognomonic for WN virus infection.
• The earliest warning in most areas was given by sightings of dead crows (Corvus brachyrhynchos - American Crow ); WN virus-positive birds were later confirmed from the same areas.
• WN virus-positive dead birds usually provided the earliest proof of viral activity in an area.
  • Data from 1999 and 2000 in the USA indicated that dead wild bird surveillance was the most sensitive method of surveillance for West Nile virus. (P39.2.w1)
  • Data from 2002 showed that in 61% of counties reporting WN virus activity (1,534 of 2,531 counties) detection of WN virus-positive dead birds was the first positive WN virus surveillance event. (D147)
• Detection of positive dead birds always preceded reporting of human cases (although knowledge of the test result did not necessarily predate onset of the human case).
  • In 527 of 589 counties (89%) in which cases of human meningoencephalitis were reported in 2002, detection of transmission of WN virus had occurred first in animals, and in 327 of the 527 counties (72%) detection of WN virus-positive dead birds was the first positive WN virus surveillance event, preceding onset of human illness by two to 252 days (median 38.5 days). (D147)
• Those counties with human cases tended to have high dead bird surveillance indices, both WN virus-positive and sightings.
  • In 2000, dead crow (Corvus brachyrhynchos - American Crow ) density in a given area was the surveillance factor most closely associated with the number of human cases in that area. There was a significant association between high dead crow density (greater than one per square mile) and human cases of WNV infection several weeks later. (J84.7.w5, J133.951.w8, J133.951.w10)
  • In 2002, many counties in which cases of WN virus infection occurred in humans tended to have high dead bird surveillance indices, both WN virus-positive birds and sightings of dead birds. (D147)
    • Counties, particularly those in midwestern states, with sparse human populations, were notable exceptions to the general tendency to have high sightings of dead birds and high numbers of WN virus-positive dead birds. (D147)
• Two-thirds of the WN virus-positive birds in New York State in 2000 were Corvus brachyrhynchos - American Crow; however individual birds of other species were the first to be confirmed WN virus-positive in 31 of 61 counties, and non-corvids in 22 counties. (J84.7.w5, J133.951.w8)
  • Testing only Corvus brachyrhynchos - American Crow would have led to a delay (days to weeks) in confirming virus presence in many counties.
• Experimental evidence of direct transmission among crows (Corvus brachyrhynchos - American Crow) (J133.951.w9) has been reported which may alter interpretation of WN virus surveillance findings if this phenomenon occurs in nature.
• The finding of a single dead virus-positive bird, even in conjunction with high infection rates in ornithophilic vector mosquitoes, at a given location early in the transmission season does not necessarily act as a reliable indicator of the level of virus transmission, and particularly the risk of human infection, later in the season. (J91.67.w1)
• Data from the northeastern USA in 2000 and 2001 indicate that counties with "high" activity of WN virus in crows early in the season (where the WN virus epizootic in crows intensified early in the season) were more likely to report human WNV infection later in the season. Two variables, (reported dead crows/area) and ((WN virus-infected crows/tested crows) x human population), were the most useful indicators of later occurrence of human disease in an area. (J279.2.w4)
• In 2001 in the USA 295/359 counties (82.2%) the earliest detection of WN virus activity was by dead bird surveillance. (P39.3.w7)
• In 2001 in Florida WN virus-positive birds provided the first confirmation of WN virus activity in 54/67 counties. 
• In New York City in 2001 avian mortality surveillance was relatively slow as a means of detecting WN virus transmission. (P39.3.w7)
• Totals of dead birds with WNV infection reported to CDC were 4305 in 2000 and 7,241 in 2001. (P39.3.w7)
• Experience in New Jersey shows that participation by members of the public reporting dead and ill birds is critical for the success of this type of surveillance. (P39.3.w18)
• Detailed data analysis using New Jersey data for 2001 showed that more useful data can be obtained if the locations of dead birds are plotted more accurately and data is not summed over large areas. (P39.3.w3)
• In Chicago in 2002 dead crow sightings were found to be correlated with later risk of human infections. (P39.4.w6)
• In Florida in 2002 dead birds, as well as sentinel chickens and infections in horses, were considered to contribute to early warning. (P48.1.w4)
• In Louisiana in 2002 findings of dead birds in specific locations occurred about two to three weeks before human cases in the same areas; dead bird surveillance was considered a useful form of surveillance. (P48.1.w3)
• Data from New York City in 2001 indicated that dead bird clusters were found in areas where later surveillance showed WN virus activity (as indicated by finding of the virus in mosquito pools and finding of a seropositive live hatch-year bird) and where later cases of human disease occurred. (J84.9.w6)
• Data from across the USA in 2002 indicated that counties with at least one WN virus-positive dead bird found early in the transmission season (by 5th August) were at greater relative risk of reporting at least one case of human WNV encephalitis or meningoencephalitis later in the year. The data was considered to suggest that in counties where there is an avian epizootic early in the transmission season, cases of WN virus-related disease in humans is more likely to occur. (J84.9.w14)

(D67, D72, D147, J84.7.w5, J84.9.w6, J84.9.w14, J91.67.w1, J133.951.w8, J133.951.w10, P39.2.w1, P39.3.w3, P39.3.w7, P39.3.w18, P39.4.w6, P48.1.w3, P48.1.w4)

• Virus has been isolated by Vero plaque assay and detected by TaqMan RT-PCR from oral swabs and cloacal swabs taken from dead birds (Corvus brachyrhynchos - American Crow, Corvus ossifragus - Fish crow and Cyanocitta cristata - Blue jay (died or euthanased following experimental WNV infection) as well as from the samples from the brains of the same birds. WN virus concentrations from all three sample types averaged greater than 10⁵ PFU (plaque forming units) by both assays. It was suggested that dead birds could be tested for WN virus by taking oral and cloacal swabs and sending the swabs, frozen, to an appropriate virological laboratory for testing. (J84.8.w7)

The following dead bird testing/submission priorities were developed for New York State and are quoted directly from the New York State West Nile Virus Response Plan - Guidance Document (D72): [Text copied directly]

NOTE: Data on WN virus in wild birds in the USA to 2007 is available summarised in map form. See: Map0001 - Spread of West Nile Virus in the USA (2000-2007) - Bird maps

Video Available: Bird Necropsy Protocol for West Nile Virus Surveillance:

• Flaviviridae- West Nile Virus (Viral Type)
  • West Nile Virus - Definitive Bird Host Species (Viral Reports)
• Mosquito Surveillance for West Nile Virus

The most suitable species (single or plural) for serosurveillance is likely to vary between geographical areas and habitat types as well as with the disease for which detection is required. Serosurveys are required to determine the optimal free-ranging bird species to use for serosurveillance. (D67)

• "The best species for serologic surveillance are those in which infection is rarely, if ever, fatal, and population replacement rates are high, ensuring a high proportion of uninfected individuals." (D67)
• Referring to data from the Czech republic, one study concluded that "in our opinion, house sparrow [Passer domesticus - House sparrow] is the most suitable species of wild birds for studying occurrence of various arboviruses in urban and suburban areas. The reasons are its high fidelity to a locality, continuous exposure to arbovirus vectors, and easy catchability of representative numbers of sparrows for testing". (J118.96.w1)
• The house sparrow (Passer domesticus - House sparrow) has been used as a sentinel for arbovirus surveillance in the USA. It is one of several suitable species of wild birds for studying the occurrence of various arboviruses in a number of urban areas, particularly in the Midwestern and Northeastern regions of the USA, but is less suitable in Florida and possibly west coastal states for SLE and WEE). (J91.32.w1, V.w42)

Protocols and specimens:

Video Available: Blood Collection Techniques for Birds:
 Internet (Web) Version (Smaller files - quicker to load)
  CD-ROM Version (Larger files - higher quality images)

• Wild birds are captured, for example using mist nets.
• A sample of blood is taken (usually from the jugular vein in passerine birds).
• Birds are individually identified (e.g. by banding with a federally authorised leg band).
  • This allows recognition of individuals if they are recaptured at a later date.
• The bird is released.
• Serum samples are tested for the presence of antibodies to WN virus.
  • For most species there will not be IgM assays therefore IgG (IgY)-detection ELISAs and PRNT (plaque reduction neutralization test) will be required. (D147)
  • The HI test is rapid and accessible but relatively low in specificity (this may be useful for screening to detect samples which may be positive to either WN virus or St Louis encephalitis virus).(P39.2.w2)
    • The HI has the advantage that species-specific conjugates are not required. (P32.1.w1)
  • ELISA suitable for multiple species have been developed. (P39.2.w2, J84.8.w9)
  • PRNT is the most sensitive and specific test but takes longer (three days) and require BSL-3 conditions as live WN virus is used in the test. (P39.2.w2)
• Individuals which are less than one year old and are WNV-antibody-positive may be considered to have been infected recently, i.e. during the current transmission season.
  • Chicks under four weeks old with weak seropositivity may have maternal antibodies.
• Individuals which are older than one year old and which are found to seroconvert provide evidence of recent transmission; collection of this data requires that the same, uniquely identified, bird is bled on more than one occasion during the transmission season. (D147)

(D67, D72, D147, J84.8.w9, J133.951.w10, P32.1.w1, P39.2.w2)

Advantages of live wild bird serosurveillance include:

• There is a long history of the successful use of wild bird serosurveillance in flavivirus surveillance.
• Local movement of resident wild birds may allow an increased chance of contact with enzootic transmission foci, thus increasing sensitivity.
  • Effectively a larger area is monitored using wild birds sampled at a given location than using captive sentinels sampled at the same location
• Set-up or maintenance costs may be minimal.
• It has a highly flexible sampling capability.
• High antibody titers are suggestive of recent infection.
• It permits evaluation of herd immunity among important amplifier hosts.
• Owner confidentiality may be less of an issue.

(D67, D72, D147, J133.951.w10)

Disadvantages of live wild bird serosurveillance include:

• More research is needed to validate the usefulness of free-ranging sentinels for monitoring WN virus transmission.
• Interpretation of serological data is complex.
  • Some types of antibody (e.g. IgM and haemagglutination inhibition antibodies) may persist in some species for only relatively short periods, so that birds may be negative on a particular serological test despite previous infection. (J133.951.w10)
  • Birds tested less than seven days after infection may not yet have detectable levels of antibodies and may therefore be seronegative. (J133.951.w10)
• Using free ranging bird surveillance requires an ability to differentiate recent infections from those acquired in previous years.
  • Birds which are less than one year old and have antibodies may be presumed to have been infected recently, within the current transmission season, although an allowance must be made for the possibility of weak seropositivity in very young birds, less than one month old, due to maternal antibodies.
  • Accurate age determination is difficult, although determination of "hatching year" versus "older" is usually possible.
  • In older birds, individuals which seroconvert from seronegative to seropositive also supply evidence of recent transmission; however this requires frequent recapture of birds during the transmission season, to gain multiple samples from individuals which have been individually and uniquely banded.
• Handling and venepuncture of reservoir species increases the risk of exposure to pathogens in feces and by accidental needle stick.
• Movement of free-ranging wild birds makes it impossible to know where the infection was acquired.
• Federal law protects most birds and possession requires state and federal permits. 
  • Banding permits require complex data reporting.
  • Some non-native species, including Passer domesticus - House sparrow, Columba livia - Rock pigeon, and Sturnus vulgaris - Common starling are exempt from federal statutes.
• Training is required for live-trapping, blood-sampling, handling and accurate determination of species and age of wild birds.
• It is generally not feasible to serially bleed individual free-ranging birds because of low recapture rates, although banding can be useful.
• Currently, serologic testing may require species-specific antiserum.
  • For most species assays detecting IgM antibodies will not be available therefore other tests, such as the plaque-reduction neutralization test (PRNT) will be required to detect WN virus-specific antibody. (D67)
  • The IgM ELISA test used for chickens does not work for wild birds. (V.w42)
• An Indirect IgG ELISA screening test has been developed and evaluated recently for the assay of antibodies to WN virus in birds. The ELISA, which used commercially available anti-wild bird horseradish peroxidase-conjugated antibodies generated from the sera of four species of bird from the orders Passeriformes (Passeriformes - Wrens, Crows, Tits & Sparrows (Order)), Columbiformes (Columbiformes - Pigeons & Doves (Order)), Galliformes (Galliformes - Pheasants, Quails & Turkeys (Order)) and Anseriformes (Anseriformes - Ducks, Geese & Swans (Order)). It was shown able to detect anti-WN virus antibodies in 23 species of birds, including both native and exotic species. All ELISA negative birds were also negative by PRNT, giving a predictive value of a negative test (PVN) 0f 100%. The predictive value of a positive test (PVP) was calculated as 70%. It was considered some ELISA-positive but PRNT-negative results might have been due to technical errors when carrying out the ELISA. The ELISA was considered to be a useful screening test, with positive results being checked by PRNT. (J84.8.w9)
• Some jurisdictions may require animal use protocols for live wild bird surveillance.

(D67, D72, J133.951.w10, P32.1.w6)

Recent experience:

• In urban epizootic transmission foci in New York City, several common species developed high seroprevalence, making them strong candidate sentinels (e.g., Passer domesticus - House sparrow, Cardinalis cardinalis - Northern cardinal, Dumetella carolinensis - Grey catbird, Zenaida macroura - Mourning dove, Columba livia - Rock pigeon), although other species may be important in other locations.
• High seroprevalence of important reservoir hosts (such as Passer domesticus - House sparrow ) in north-eastern Queens in 1999 preceded low transmission activity to humans in the same neighbourhoods in 2000.
• In New York State: "Serologic testing of apparently healthy wild birds has indicated low WNV positivity rates early in the mosquito transmission season and higher positivity rates later in the season. Thus, it does not appear at this time that monitoring antibody levels in wild-caught healthy birds will provide an early warning of viral activity, in comparison to dead bird surveillance. However, as with sentinel flocks or groups, further research is recommended to understand infection rates, clinical impact, and the potential surveillance role of wild bird serologic testing. Priority may be considered for monitoring antibody levels of wild caught healthy crows in order to help determine the case fatality rate from WNV. Although various agencies will consider this type of monitoring for research purposes in New York State in 2001, it is not recommended as a core surveillance component for LHUs [Local Health Units]." (D72)
• In New York City in 2000,  3/440 (0.7%) of wild house sparrows (Passer domesticus - House sparrow ) and 1/150 (0.7%) other birds (a mockingbird (Mimus polyglottos - Northern mockingbird)) tested seropositive for WN virus. (P39.2.w2)
• Several seropositive hatching-year wild birds were identified in New York City in 2000. (J133.951.w10)
• A study in 2001, testing birds at three locations where single dead WN-virus positive birds had been found early in the transmission season in 2000, found neutralisation antibodies (detected by PRNT) in a hatch-year Passer domesticus - House sparrow (one of 85 tested) at one of the locations (Floral Park). Of other birds tested, 25/75 were seropositive at Floral Park. For the two other sites, 0/82 hatch-year sparrows and 1/24 other birds were seropositive at New Rochelle while 0/76 hatch-year sparrows and 0/47 other birds were seropositive at Babylon; WN virus-infected mosquitoes were found at all three sites. (J91.67.w1)
• In Illinois in 2001 sera were collected from 7,786 birds in areas where dead birds had tested positive for WN virus and none were positive; it was concluded that wild bird serology using the IgM ELISA appeared not to be useful for WN virus surveillance in Illinois. (P39.3.w20)
• In New York City, house sparrows (Passer domesticus - House sparrow) appear to be a suitable target species for live wild sentinel bird surveillance. (P39.2.w2)
• In New York City in transmission foci high seroprevalence was noted in several common species, which would therefore be strong candidates for sentinels; it is recognised that in other locations different species may be important as sentinels. (D147)
• In several regions it has been noticed that this surveillance system requires a much greater effort than do other surveillance systems. (D147)

(D67, D72, D147, J91.67.w1, J133.951.w10, P39.2.w2, P39.3.w20)

Information required to maximise the effectiveness of live wild bird serosurveillance include:

• Locations of WN virus enzootic transmission foci in an area.
• Which species are most frequently infected within such enzootic transmission foci.
• The species acting as important reservoir hosts in a given area, which may then be targeted for surveillance efforts.
• The ability of different species to develop antibodies which are detectable by available serological tests, the time after infection at which antibodies appear and the length of time for which such antibodies persist.

(J133.951.w10)

Video Available: Blood Collection Techniques for Birds:
  Internet (Web) Version (Smaller files - quicker to load)
  CD-ROM Version (Larger files - higher quality images)

• Regional Assessment of West Nile Virus
• Clinical Approach to Cases for Veterinary and Medical Personnel
• Flaviviridae- West Nile Virus (Viral Type)
  • West Nile Virus - Definitive Bird Host Species (Viral Reports)
  • West Nile Virus - Detection and Identification Techniques (Viral Reports)
• West Nile Virus Disease (Viral Disease)
  • West Nile Virus Disease - Editorial Overview of Disease Characteristics for specified SPECIES-TAXA

The use of sentinel chickens for urban and state-wide arbovirus surveillance (Florida and California) has been successfully used for St Louis encephalitis virus (SLE) and Western equine encephalitis virus (WEE) surveillance (J84.7.w22, J91.46.w1, V.w42) and has recently been successful for WN virus in Florida. (W197.Dec02.WNV1)

Protocols and specimens:

Video Available: Blood Collection Techniques for Birds:
 Internet (Web) Version (Smaller files - quicker to load)
  CD-ROM Version (Larger files - higher quality images)

"The ideal sentinel bird is a species that is uniformly susceptible to infection, is resistant to disease, rapidly develops a detectable immune response, is easily maintained, presents negligible health risks to handlers, does not contribute to local pathogen transmission cycles, and seroconverts to the target pathogen prior to the onset of disease outbreaks in the community. There is probably no ideal sentinel species for any zoonotic pathogen." (J133.951.w10)

Captive birds of a species which is known to become infected with WN virus and to develop antibodies in response to infection, but which does not show clinical illness or mortality in response to infection, may be used as sentinels.

• The sites at which sentinel flocks are to be kept are determined.
• The species to be kept as sentinels is chosen.
• The type of cage/enclosure to be used is chosen.
• Flocks of birds seronegative for WN virus are set up in cages/enclosures and maintained for the duration of the mosquito season.
• Birds in sentinel flocks are blood sampled regularly, for example every two weeks.
  • Testing every 7-10 days has been recommended to minimise the chance of false positives due to "missing" an IgM response. (J133.951.w10)
• Sera from blood samples are tested for WN virus antibodies.
  • Serum may be screened using haemagglutination inhibition (after appropriate treatment for removal of non specific inhibitors from avian serum), ELISA or PRNT.  (D147)
  • Neutralisation must be used to confirm positive tests in order to rule out false positives and cross-reactions due to infection with related flaviviruses such as St. Louis encephalitis virus (SLEV). (D147)
• Any individual which develops antibodies for WN virus is replaced by a new seronegative individual.

(D67, D72, D147, J133.951.w10)

• Candidate sentinel bird species which have been shown to be incompetent vectors for WN virus include the domestic chicken, ring-necked pheasant (Phasianus colchicus - Common pheasant), bobwhite quail, Japanese quail and feral pigeon (Columba livia - Rock pigeon). (P39.3.w7)

Advantages of captive sentinel bird surveillance:

• There is a long history (more than sixty years) of the successful use of chickens in flavivirus (other than WN virus) surveillance.
• Birds are readily fed upon by vector mosquitoes.
• Captive birds can be serially bled, and the geographic location of infection is not in question.
• The system is flexible and therefore can be expanded and contracted as appropriate.
• Mosquito-abatement districts, if established, can bear the responsibility for maintaining flocks, bleeding birds, and submitting specimens for testing.
• Laboratories with the ability to charge for testing may defray the expenses of testing by charging nominal fees per test.
• Captive sentinels are useful for monitoring WN virus activity in transmission foci. (P39.3.w7)
• Compared to free-ranging live bird surveillance collection of specimens is inexpensive. (D147)

(D67, D72, D147, J133.951.w10, P39.3.w7)

Disadvantages of captive sentinel bird surveillance:

• Sentinel flocks detect only focal transmission, requiring that multiple flocks be positioned in representative geographic areas. This is particularly true when vector mosquitoes have short flight ranges (for example Culex pipiens (Culex pipiens complex - Northern and Southern house mosquitoes)).
  • Lack of knowledge of the location of enzootic transmission foci makes choosing appropriate sites for locating sentinels problematic. (J133.951.w10)
• Flocks are subject to vandalism and theft, limiting usefulness in urban areas.
• Flocks must be protected from predators.
• Costs of setting up and maintaining flocks may be considerable. This includes purchase of initial and if necessary replacement birds, provision of appropriate, predator-proof accommodation, protection from human thieves or vandals, feeding etc.
• Training is required for proper maintenance and sampling of flocks, including for repeated blood sampling of birds with minimal risk of injury to the birds or personnel.
• Pre-existing flocks may already be exposed due to previous local WN virus transmission.
• More research is needed to validate the usefulness of sentinel captive birds for monitoring WN virus transmission.
• There is a time lag between the occurrence of infection in a sentinel and detection of infection in the bird, consisting of the time from infection to the development of detectable levels of antibodies, the time from the development of such antibodies to the next blood sampling date, the time taken for samples to reach the laboratory, time for the tests to be carried out to produce results and time for results of tests to be communicated.
• Sentinel chickens, if infected with WN virus, sometimes shed small amounts of virus in their droppings. This represents a small potential zoonotic risk to personnel.
• Sentinel birds such as chickens may develop viraemia sufficient to infect mosquitoes.

(D67, D72, D147, J133.951.w10)

Recent Experience:

• Experimental studies have shown that chickens Gallus gallus domesticus, pigeons (Columba livia - Rock pigeon) and pheasants (Phasianus colchicus - Common pheasant) are safe sentinels. However, small amounts of WN virus have been isolated from cloacal specimens, and thus virus may be shed in the feces. There is no evidence to date that this route of exposure has served as a source of human infection.
• Field studies of avian seroprevalence in Queens in 1999 indicated that captive chickens were frequently infected. In Staten Island in 2000, captive pigeons were frequently infected.
• Prior to 2002 in the USA detection of WNV infection in sentinel chicken flocks was not a sensitive method for detection of WN virus activity in an area however during 2002 the use of sentinel chickens appears to have been more effective.
  • Chickens were not effective sentinels in the northeastern USA in 2000. (J214.267.w11)
  • Chickens were used as sentinels in 2000 in selected counties in New York State, New York City, New Jersey, Pennsylvania, Maryland Delaware. Small numbers of seroconversions were detected late in the season in New Jersey, New York City and New York State, after the onset of human cases of WNV infection. Of more than one hundred sentinel chickens maintained in various locations in New York State, eight were positive for WN virus.(J133.951.w10)
  • During 2000, in New York City 13 coops of seven birds each were placed at live poultry markets in four boroughs and a further flock of 15 birds was placed on Staten Island. Seven of the 106 birds seroconverted and one dead chicken tested positive for WN virus. In New York State flocks were placed in four counties, with 103 birds in Suffolk, 24 in Rockland, 70 in Westchester and 47 birds in Onondaga county. One of the 244 birds seroconverted (in Westchester on August 4th) (IgG ELISA, confirmed by PRNT). In New Jersey one to two flocks of three birds each were maintained in all 21 counties, with more flocks in two counties; four of the 102 birds seroconverted (HI test, confirmed by PRNT). (P39.2.w2)
  • During 2001 in the USA detection of WNV infection in sentinel chickens occurred in five chickens in New Jersey, 170 chickens in Florida and three in North Carolina. (W27.12Dec01.wnv1)
  • In 2001 in the USA in 6/359 counties (1.7%) the earliest detection of WN virus activity was by sentinel chicken surveillance. (P39.3.w7)
    • Detection of seroconversions in sentinel chickens rose from just 13 in 2000 to 218 in 2001. (P39.3.w7)
  • In New York City in 2001 sentinel chickens and pigeons were both found to be useful for monitoring WN virus transmission in enzootic study sites. (P39.3.w7)
  • During 2002, seroconversions were reported in 96 sentinel chicken flocks in Florida, Nebraska, New York City and Pennsylvania. (W27.01Sept02.wnv1)
  • During 2002 in Florida, up to16th December, positive samples had been detected in 957 sentinel chickens from 29 counties. (W197.Dec02.WNV1)
  • During 2002 in Florida, seroconversion in sentinel chickens was one criterion for placing a county on medical alert. There were a total of 1,021 cases in chickens during the year. Dead birds and infections in horses were also considered to contribute to early warning. (P48.1.w4)
  • In New York State some mortality in chickens was attributed to WN virus infection. (D147)

(D67, D72, D147, J133.951.w10, P39.2.w2, P39.3.w7, P48.1.w4, W27.12Dec01.wnv1, W27.01Sept02.wnv1, W197.Dec02.WNV1)

Factors which may affect the success of captive sentinel bird surveillance:

• Determination of enzootic transmission foci at which sentinel flocks should be placed (this may require collection of several years of data).
• Placement of flocks within microhabitats conducive to attack by mosquitoes.
• Cage designs.
• Cage height above ground level.
• Flock size.
• Relative attractiveness of sentinel species to the mosquito vectors.
• (J133.951.w10)

NOTE: Data on surveillance using sentinel flocks in the USA to 2007 is available summarised in map form. See: Map0001 - Spread of West Nile Virus in the USA (2000-2007) - Sentinels maps

Video Available: Blood Collection Techniques for Birds:
  Internet (Web) Version (Smaller files - quicker to load)
  CD-ROM Version (Larger files - higher quality images)

• Regional Assessment of West Nile Virus
• Clinical Approach to Cases for Veterinary and Medical Personnel
• Flaviviridae- West Nile Virus (Viral Type)
  • West Nile Virus - Definitive Bird Host Species (Viral Reports)
  • West Nile Virus - Detection and Identification Techniques (Viral Reports)
• West Nile Virus Disease (Viral Disease)
  • West Nile Virus Disease - Editorial Overview of Disease Characteristics for specified SPECIES-TAXA

Proposed zoo-based surveillance project in the USA:

The proposed project is divided into two phases: [Text copied directly]

(W253.Nov01.WNV1)

Samples to be taken:

Video Available: Blood Collection Techniques for Birds:
 Internet (Web) Version (Smaller files - quicker to load)
  CD-ROM Version (Larger files - higher quality images)

These include blood samples, tissue samples taken at necropsy and cloacal/nasopharyngeal swabs.

• Blood samples: 0.5ml of citrated (preferred choice) or heparinised whole blood, or 0.25ml heparinised plasma (from birds) or serum (from mammals).
• Tissue samples: minimal submission to include sections of brain, heart, and kidney, stored at -70°C or in liquid nitrogen or if sent for testing within 2-3 weeks, at -20°C.
• Cloacal/nasopharyngeal swabs: placed in viral transport media and shipped on ice packs, or, kept at 70°C and shipped on dry ice. (W253.Nov01.WNV1, W253.Nov01.WNV2).

Advantages of zoo-based surveillance:

• Zoo-based surveillance provides a unique opportunity for detection of both clinical infections and seroconversion in a wide variety of native and non-native species because illness, even slight, is more likely to be noticed in zoo animals than in free-living wild animals and because frequently sequential serum samples are available from positively identified individuals.
• Additionally, it is possible in zoos to study survivors of WNV infection to determine any long term effects of infection; this is not normally practical with wild animals. 

(J4.219.w3)

Disadvantages of zoo-based surveillance:

• Zoo-based surveillance is necessarily limited in location to sites at which a zoo is present. (V.w5)

Recent experiences of zoo-based surveillance:

• During the period 1st August 2001 to 28th February 2002 zoo-based surveillance tested more than 1,450 animals (967 birds of 195 species, 40 equids of eight species, 436 other mammals of 110 species and 20 reptiles of sis species) from 64 participating institutions in 30 states (including the District of Colombia) and found 30 WN virus antibody-positive birds, 29 WN virus-positive birds, one antibody-positive reptile and a further 13 antibody-positive animals (10 birds, two mammals and one reptile) awaiting confirmation following initial screening. Antibody-positive animals were found in zoos from nine states and virus-positive animals were found in zoos from four states. No positive zoo samples were the first WN virus activity-confirmatory event in any area. Six of the 30 confirmed seropositive animals (20%) presented with neurological signs while the remaining 24 were found by serosurvey. Six of the 29 virus-positive captive animals (21%) showed clinical signs of illness prior to death; the remaining 23 virus-positive individuals were found dead within their enclosures or in the zoo grounds. Four captive birds were detected virus-positive with PCR but the infection was not confirmed by virus isolation. One captive bird was virus-positive but antibody-negative. For seven animals initial screening tests were WN virus antibody-positive but end-point titration of PRNT did not allow definitive classification of WN virus or St. Louis encephalitis infection. Six birds which had been shipped to one institution from Africa were found to have weakly positive WN virus antibodies in serum samples which had been taken and archived during their period in quarantine. (P39.3.w8)

Video Available: Blood Collection Techniques for Birds:
 Internet (Web) Version (Smaller files - quicker to load)
  CD-ROM Version (Larger files - higher quality images)

Video Available: Bird Necropsy Protocol for West Nile Virus Surveillance:
Internet (Web) Version (Smaller files - quicker to load)
CD-ROM Version (Larger files - higher quality images)

• Regional Assessment of West Nile Virus
• Clinical Approach to Cases for Veterinary and Medical Personnel
• Flaviviridae- West Nile Virus (Viral Type)
  • West Nile Virus - Definitive Bird Host Species (Viral Reports)
  • West Nile Virus - Definitive Mammal Host Species (Viral Reports)
  • West Nile Virus - Definitive Amphibian and Reptile Host Species (Viral Reports)
  • West Nile Virus - Detection and Identification Techniques (Viral Reports)
• West Nile Virus Disease (Viral Disease)
  • West Nile Virus Disease - Editorial Overview of Disease Characteristics for specified SPECIES-TAXA

Testing of mammals does not appear to be essential for WN virus surveillance, because WN virus positive mammals do not provide the earliest indication of viral activity when dead bird surveillance is already established. (D72)

Testing of mammals can be problematic in regard to both test interpretation and resources available for testing.

• Testing of mammals may be indicated as a service to owners and veterinarians, or for research purposes. (D72)
• It is important to consider alternative causes of encephalitis in mammals, for example in New York State, mammals that have died of encephalitis are more likely to have died of rabies than WNV infection and rabies can be transmitted to people before those infected mammals die; therefore it continues to be essential that all mammals with neurological signs, that have had contact resulting in possible rabies exposure to people, pets, or domestic animals, be submitted for rabies testing. (D72)
  • Rabies and WNV testing of mammals are not exclusive; both can be performed on the same suspect animal but rabies testing should be conducted first and determined to be negative before the brain specimens is tested for WNV (however, immunohistochemistry could be performed on rabies positive brains that have been formalin fixed). (V.w42)

• Flaviviridae- West Nile Virus (Viral Type)
  • West Nile Virus - Definitive Mammal Host Species (Viral Reports)
  • West Nile Virus - Detection and Identification Techniques (Viral Reports)
• West Nile Virus Disease (Viral Disease)
  • West Nile Virus Disease - Editorial Overview of Disease Characteristics for specified SPECIES-TAXA

In New York State, "as a service to veterinarians and their clients, and to help advance research into WNV’s impact on animals, Cornell University [W275] will continue to provide WNV testing for ill owned animals, as resources permit." (D72)

• Rabies and WNV testing of mammals are not exclusive; both can be performed on the same suspect animal but rabies testing should be conducted first and determined to be negative before the brain specimens is tested for WNV (however, immunohistochemistry could be performed on rabies positive brains that have been formalin fixed). (V.w42)

• Regional Assessment of West Nile Virus
• Clinical Approach to Cases for Veterinary and Medical Personnel
• Flaviviridae- West Nile Virus (Viral Type)
  • West Nile Virus - Definitive Bird Host Species (Viral Reports)
  • West Nile Virus - Definitive Mammal Host Species (Viral Reports)
  • West Nile Virus - Definitive Amphibian and Reptile Host Species (Viral Reports)
  • West Nile Virus - Detection and Identification Techniques (Viral Reports)
• West Nile Virus Disease (Viral Disease)
  • West Nile Virus Disease - Editorial Overview of Disease Characteristics for specified SPECIES-TAXA