Guidelines on Wild Bird Surveillance for Highly Pathogenic Avian Influenza H5N1 Virus

[Sally Furniss]

Guidelines on Wild Bird Surveillance for Highly Pathogenic Avian Influenza H5N1 Virus

<nobr>Vittorio Guberti²</nobr> and <nobr>Scott H. Newman³^,4^,5</nobr>

² Istituto Nazionale Fauna Selvatica, Via Ca’ Fornacetta, 9-40064 Ozzano E. (BO), Italy
³ United Nations Food and Agriculture Organization, Via delle Terme di Caracalla, 00100 Rome, Italy
⁴ Wildlife Conservation Society, Field Veterinary Program, 2300 Southern Blvd., Bronx, New York, 10460 USA
⁵ Corresponding author (email: Scott.Newman@fao.org<script type="text/javascript"><!-- var u = "Scott.Newman", d = "fao.org"; document.getElementById("em0").innerHTML = '<a href="mailto:' + u + '@' + d + '">' + u + '@' + d + '<\/a>'//--></script>)

<!-- ABS --> ABSTRACT: The recent spread of highly pathogenic avian influenza (HPAI)H5N1 from Southeast Asia into Europe and Africa emphasizes theneed to better understand the mechanisms by which the diseaseis spread. Current wildlife surveillance has been limited toa combination of targeted sampling, opportunistic sampling,and recovery of dead wildlife. Recent interest in avian influenzasprovides an opportunity to develop a global surveillance programfor diseases in wild birds. This program should be based onspecific principles with clearly defined aims, sound epidemiologicalsampling justifications, and sufficient technical skills andcapabilities for appropriate field and laboratory activities.At present, although detection of H5N1 in healthy wild birdsis sporadic, surveillance programs should be focused on: 1)determining the role wildlife plays in the epidemiology of HPAIH5N1 virus, and 2) guiding strategies to prevent disease exposureto humans and poultry. Surveillance should incorporate activeand passive components using available natural history informationsuch as intra- or interspecies mingling, population size, migratorypatterns, and seasonal large-scale aggregations. Passive surveillancecan include samples from bird banders/ringers, other wild birdresearch efforts, oil-spill response efforts, rehabilitationcenters, zoological collections, beached-bird surveys, and mortalityevents. A valuable wildlife surveillance program based on ecologicaland epidemiological information will require large-scale collaborationamong national governments and ministries, multilateral agencies,nongovernment organizations, academicians, veterinarians, virologists,ornithologists, and many others.
Key words: Avian influenza, guidelines, HPAI, H5N1, surveillance, waterfowl, wild birds.
<!-- null --> Over the past 5–10 yr, increasing numbers of infectiousdiseases in humans, wildlife, and domestic animals have emerged,and diseases such as mad cow disease, Lyme disease, monkey pox,nipah virus, and severe acute respiratory syndrome (SARS) havebecome common household terms. Most recently, with the emergenceof highly pathogenic avian influenza (HPAI), subtype H5N1, globalattention increased as people were informed that the virus couldpotentially represent the source of the next human influenzapandemic (de Jong et al., 1997).
While the scientific community acknowledges that this is primarilya poultry disease, and that prevention and control measuresshould be addressed at the level of agricultural productionand farming practices, concern has been raised about the rolethat wild birds may play in harboring, perpetuating, and transmittingavian influenzas to new geographic locations internationallyor intercontinentally. Wildlife species were not implicatedas reservoirs or carriers of HPAI H5N1 during the initial outbreaksin poultry in Southeast Asia. From 2003 to 2005, however, large-scalemortality events in wild birds in China, Russia, Mongolia, andEastern Europe clearly indicated that wild birds were susceptibleto HPAI H5N1 (EMPRES Watch, 2005, 2006a; OIE, 2005–2006).Although suggestive, these initial events did not provide compellingevidence that infected wild birds could move these viruses overlong distances during migration.
As of August 2006, wild bird mortality related to HPAI H5N1had been documented in Azerbaijan, Hungary, Romania, Serbia,Turkey, and Ukraine, where concurrent poultry outbreaks wereoccurring. In other countries, such as Bosnia-Herzegovina, Bulgaria,Croatia, Czech Republic, Georgia, Greece, Poland, Slovakia,and Slovenia, wild bird mortality occurred independent of knownconcurrent poultry outbreaks. This suggested that wild birdswere infected elsewhere and flew into these countries beforedying (EMPRES Watch, 2006b; Gilbert et al., 2006). This possibilitywas reinforced from fall 2005 through winter 2006 when HPAIH5N1 mortality in wild birds was documented across Western Europe.This mortality was mainly observed in mute swans (Cygnus olor),but other species were involved including mallard (Anas platyrhynchos)and common pochard (Aythya ferina) in France, great crestedgrebe (Podiceps cristatus) in Spain, little grebe (Tachybaptusruficollis), tufted duck (Aythya fuligula), and common pochardin Switzerland, and whooper swan (Cygnus cygnus) in the UnitedKingdom (OIE, 2005–2006; USGS, 2006).
Although it had not yet been demonstrated that wild birds shedthe virus during long-distance movements, a potentially importantmechanism of disease spread via wildlife movements needed tobe examined more carefully. To address the role of wild birdsin the possible spread of HPAI H5N1, wildlife surveillance programsin 2005–2006 included sampling in the European Communityin the framework of a European Commission (EC) activity; EasternEurope, Asia, and Africa by the Food and Agriculture Organization(FAO) of the United Nations, Centre de Coop&#233;ration Internationaleen Recherche Agronomique pour le D&#233;veloppement (CIRAD),and Wetlands International; Canada by the Canadian WildlifeService (CWS); United States by the US Fish & Wildlife Service(USFWS) and US Geological Survey (USGS); South and Central America,Africa, Eastern Europe, and Asia by the Wildlife ConservationSociety’s Global Avian Influenza Network for Surveillance(GAINS); and many national programs, including, but not limitedto, China, Russia, Mongolia, Switzerland, Germany, Vietnam,Indonesia, and Iran. These surveillance programs were specificallyfocused on surveillance for HPAI H5N1.
To date, over 80 species from 14 orders of birds have been foundpositive for HPAI H5N1 (USGS, 2006); almost all positive sampleswere from moribund, sick, or dead birds. These birds generallyfell into three categories: migratory waterfowl, bridge species(most likely nonmigratory) which may transport disease frompoultry to wildlife or vice versa, and predatory birds whichlikely scavenged dead birds (poultry or wild birds) or predatedsick birds.
Information from healthy migratory birds includes more than100,000 samples from EC, over 5,000 samples from Eastern Europe,Asia, and Africa, over 20,000 samples from the Americas, andhundreds of national program samples. Only one study thus farhas demonstrated apparently healthy waterfowl positive for HPAIH5N1. The six positive birds (ducks) were not identified specificallyin over 4,000 birds sampled in China from an area also usedto raise domestic ducks (Chen et al., 2005).
In general, surveillance programs are difficult, complicated,and require a large, coordinated planning team(s). Most importantly,surveillance programs must be targeted and focused in orderto obtain valuable information about disease ecology. Successfulprograms require biologists/ornithologists trained in propermonitoring and sample-collection techniques, year-round monitoringat wildlife habitat where species of concern are known to aggregate,and collaboration among biologists/ornithologists, veterinaryservices, and chief veterinary officers so coordinated diseasesampling and testing activities can take place.
Broadly defined, surveillance includes the systematic collectionof data referring to a particular disease in a population, theevaluation of such data, and the dissemination of the derivedinformation to all who need to know. The classical definitionencompasses two approaches and when applied specifically toHPAI H5N1 includes passive surveillance (opportunistic samplingfrom multiple sources especially related to wild bird morbidityand mortality events) and active surveillance (targeted samplesto detect the targeted disease or agent).
Passive sampling takes advantage of pre-existing programs orsources of wild birds, such as rehabilitation or zoologicalexhibit animals, beached-bird monitoring programs, or birdscollected through properly managed hunting programs. Samplescannot be guaranteed from a predetermined number of individuals,nor a specific species or group of species. In contrast, activesurveillance programs can guarantee a certain level of detectionand predetermined level of accuracy.
The most important goal of an HPAI H5N1 wildlife surveillanceprogram is early detection of the virus, usually from a speciesor group of species that serve as a potential epidemiologicalreservoir. In the case of HPAI H5N1, a wildlife reservoir hasnot been identified or confirmed, and this limits the classicalsampling approach according to a bimodal distribution (positiveand negative individuals) in an infected population. Becausean expected prevalence is not known, sampling intensity shouldconservatively focus on finding at least one positive animalassuming a very low prevalence (i.e., 0.5&#37. This often requiresvery large sample sizes.
The number of samples taken is a function of the populationsize combined with the expected number of positive animals.For infectious diseases, the population can be defined as thegroup of animals that has the same probability of being positive(or negative) during the surveillance period. Also, the geographicalscope is of paramount importance: should surveillance occurat every lake or wetland in each country or at selected wetlandsthat have certain characteristics? The critical community size(CCS), defined as the number of individuals able to maintainan infection in the environment for a certain period of time,should guide this decision when possible. Although not knownfor HPAI H5N1, the CCS for low pathogenic avian influenza (LPAI)in wintering dabbling ducks indicates that wetlands containing1,200–1,500 waterfowl should be included in a surveillanceprogram (Guberti et al., 2007). The number of wetlands mightbe reduced by combining population size with other risk factors,such as presence or number of backyard poultry, commercial farms,rice fields, or integrated farms. Considering the mortalityassociated with HPAI H5N1 in some wild species (Hulse-Post et al., 2005;Sturm-Ramirez et al., 2005; Brown et al., 2006), the CCS forthe virus must be higher than that required for LPAI viruses.
Finally, surveillance strategies for HPAI H5N1 should considerthe potential temporal and spatial factors that affect prevalenceas it exists with LPAI in wild bird populations. In ducks inthe Northern Hemisphere, prevalence is highest during late summerand early fall (premigration staging) and declines rapidly duringwinter (Hinshaw et al., 1980; Webster et al., 1992). If HPAIH5N1 virus prevalence trends are similar to LPAI viruses, theprobability of finding the virus in northern latitudes wouldbe highest during the premigration staging and molt period atthe end of summer; likewise, the virus may be more difficultto detect as these birds migrate south (e.g., sub-Saharan Africaor Mediterranean Basin), especially at overwintering sites.
Bearing in mind that relevant information (CCS, expected prevalence,metapopulation sizes, species susceptibility and lethality)is still lacking for the establishment of an HPAI H5N1 surveillanceprogram, and that logistical and financial limitations exist,we recommend a combined passive and active surveillance scheme.This will provide the best approach to detect HPAI H5N1 earlyand to assess viral presence in potentially endemic areas. Thecombination of active and passive surveillance will provideanswers to the following important questions: do HPAI H5N1 virusesexist and persist in wild bird populations, and if so, where,when, and in what species do they occur? To answer these questions,the following recommendations are offered.

1. Implement bird observation and monitoring programs at key wildlifehabitats to enable daily to weekly monitoring of bird deathsand facilitate collection and testing of fresh carcasses. Thismonitoring requires minimally trained staff, a reporting mechanismand information repository, and coordination with veterinaryservices to provide diagnostic support. This level of surveillancecan be applied easily to any important bird habitat. Supplementaltraining of personnel involved in monitoring programs couldprovide insight into conservation issues such as species presenceand absence during the year, numbers of individuals, behavioralecology, changing environmental conditions, and presence ofother avian diseases. If financial resources are limited, effortshould be focused on sick or dead birds to facilitate findingHPAI H5N1 at a few select habitats where additional risk factorsexist.
2. Based on the international epidemiological situationof HPAIH5N1, a prompt investigation of wild bird morbidityor mortalityevents should be considered at local areas whenat risk. Inareas that are ecologically linked with recent outbreaksorongoing poultry outbreaks, single cases of mortality in high-riskspecies (mainly Anseriformes and Charadriiformes) must be investigated.In low-risk species and when concurrent poultry outbreaks arenot occurring, unusual mortality events involving 10 or moreindividuals in less than a week within a 10 km radius from awetland, lake, or pond should be investigated. The costs associatedwith this type of response are relatively minor, and informationgained about other infectious diseases will be inherently valuable.If funding is limited, diagnostic work should be restrictedto check for HPAI H5N1 virus only. Finally, full understandingof the disease ecology of HPAI in endemic or outbreak areasneeds to include potential transmission between agriculturalproduction and wetland habitats by "bridge species" such asdoves (Columbidae), starlings (Sturnidae), sparrows (Passeridae),swallows (Hirundinidae), and other primary synanthropic birdspecies. An effective communication network to report any epidemiologicalevent linked to the possible spread of the virus needs to bedeveloped.
3. If expertise in bird capture and handling exists,and financialsupport is available, bird surveillance shouldbe conductedon healthy, live birds using the principles outlinedin theprevious section, including: a viral prevalence lessthan 1%;tracheal swabs (or oropharyngeal swabs in smaller species)togetherwith cloacal swabs collected on all birds; sample collectionduring seasons when viral detection is most likely; if waterfowlare to be sampled, wetlands containing more than 1,200–1,500waterfowl should be targeted; and if possible, habitats adjacentto poultry farms, or species visiting these farms, should besampled.

 If capture of live healthy birds is not an option, then samplesshould be collected from nationally sanctioned hunted birds;however, prevalence may be affected if hunting seasons occurin late fall or winter. Note that this strategy lacks knowledgeabout the CCS and the populations investigated. Also, financiallimitations and small sample sizes often associated with livebird captures may dictate sampling and laboratory testing efforts.
Sampling efficacy can be significantly improved if a wild reservoirof HPAI H5N1 virus exists and is identified and when globalcoordination and information exchange can address macroareasand species to be sampled in light of the ongoing spatial distributionand epidemiological situation of HPAI H5N1 virus. Such informationwill improve surveillance at the local scale, especially inselecting species at risk, appropriate sample sizes, and appropriateseason. Current HPAI H5N1 virus surveillance strategies usingmany locations, each with a relatively small sampling intensity,may not reveal infected birds because the prevalence (as of1 April 2007) appears to be very low, and sampled individualsmay not belong to the same population.
Only regular and repeated surveillance of appropriate habitatswill have the sensitivity to detect mortality and provide opportunitiesto sample moribund or fresh carcasses where HPAI H5N1 is mostlikely to be found. Other ongoing programs, such as beached-birdmonitoring, wildlife rehabilitation, or bird banding/ringing,also may provide inexpensive samples as long as the speciesare appropriate for HPAI H5N1 testing. These "passive surveillance"samples are opportunistic and capitalize on previously establishedprograms.
Active targeted sampling programs developed by the EC, FAO,CWS, USGS, USFWS, and GAINS are only possible with significantfinancial backing. However, no active surveillance programscan guarantee early detection of the HPAI H5N1 virus if infectionis sporadic and prevalence is low in wild birds. Surveillanceprograms can detect diseases only when a detectable level ofdisease is present, and even then, success depends on a combinationof factors including viral epidemiological parameters, hostpopulation size, and sampling intensity.
Understanding whether a wild reservoir species exists, and thenidentifying possible drivers of HPAI emergence, such as anthropogenicchanges to the environment, intensified agricultural practices,or selection pressure for dominance of viral strains adaptedto survive in a modified environment (Morse, 1993; Schragg and Wiener, 1995),are important next steps in the control and management of HPAIH5N1 virus. Environmental, agricultural, and ecological conditionsalso affect disease infection rates, likelihood of transmission,and maintenance or persistence of disease. Once risk factorsare clearly identified, resource management decisions, includingthose specific for improved biosecurity, can minimize diseasespread among domestic poultry, wildlife, and humans.

Key words: Avian influenza, guidelines, HPAI, H5N1, surveillance, waterfowl, wild birds.
<!-- FN --> <!-- null --> ¹ Presentation at the FAO and OIE International Scientific Conferenceon Avian Influenza and Wild Birds, Rome, 30 and 31 May 2006

http://www.jwildlifedis.org/cgi/content/abstract/43/3_Supplement/S29

[HenryN]

Re: Guidelines on Wild Bird Surveillance for Highly Pathogenic Avian Influenza H5N1 Virus

Foxes (all the usual suspects) guarding the chickens.

[Gert van der Hoek]

Re: Guidelines on Wild Bird Surveillance for Highly Pathogenic Avian Influenza H5N1 Virus

Just a question: eggs can contain H5N1

What about eggs of wild birds? Especially birds like ducks, geese, crow-like birds?

Did not see any research on that? May be monitoring the eggs?



Detection of Hong Kong 97-like H5N1 influenza viruses from eggs of Vietnamese waterfowl.

Li Y, Lin Z, Shi J, Qi Q, Deng G, Li Z, Wang X, Tian G, Chen H.
Animal Influenza Laboratory of the Ministry of Agriculture Sciences and National Key Laboratory of Veterinary Biotechnology, Harbin Veterinary Research Institute, CAAS, Harbin, PR China.

Three H5N1 influenza viruses were isolated from shell washes of duck and goose eggs confiscated from travelers coming from Vietnam. All eight gene segments of these viruses share high sequence identity with the H5N1 avian influenza viruses that caused outbreaks in poultry and humans in Hong Kong in 1997. Animal studies indicate that these isolated viruses are able to replicate in mouse lung and could be found in the organs of ducks without causing any clinical signs or death. However, the viruses are highly pathogenic for chickens. Although the source of these recently isolated Hong Kong 97-like H5N1 viruses is undetermined, their detection in the egg shell of duck and goose suggests that this particular genotype of H5N1 virus may have re-emerged in nature or may have been circulating continuously.

http://www.ncbi.nlm.nih.gov/pubmed/16532371

Abstract:

Avian influenza virus (AIV) was recovered from the internal contents of eggs, including mixture of albumen and allantoic fluid, and from the oviduct of naturally infected Japanese quail (Coturnix coturnix japonica) flocks in the southern part of Thailand. The virus titers of 104.6-106.2 ELD50/mL were directly measured from the internal content of infected eggs. The virus was isolated by chorioallantoic sac inoculation of embryonating chicken eggs. Infected allantoic fluid was identified as hemagglutinating virus and then was indicated the presence of H5 hemagglutinin. The virus was confirmed to be H5N1 subtype influenza A virus by reverse transcriptase-polymerase chain reaction. Additionally, real-time reverse transcriptase-polymerase chain reaction assay could specifically detect influenza virus subtype H5. Furthermore, indirect fluorescent antibody (IFA) test by using specific anti-influenza A monoclonal antibody indicated that virus antigens were detected in the parenchyma of multiple tissues. Systemic localization of viral antigen detected was certainly considered to be viremic stage. In addition, influenza virus antigen was also detected by IFA in allantoic fluid sediments isolated from internal content of egg or oviduct. The conclusion of isolated AIV type A subtype H5N1 from these two infected materials was correlated to the viremic stage of infection because the virus antigens could be observed in almost all tissues. Conclusively, the need for adequate safeguards to prevent contamination and spread of the virus to the environment during movement of eggs—including hatching eggs, cracked eggs, and other relevant infected materials— or egg consumption from area of outbreak is emphasized and must not be ignored for the reasons of animal, public, and environmental health.

http://www.blackwell-synergy.com/doi/abs/10.1196/annals.1373.020?cookieSet=1&journalCode=nyas