Global Patterns of Influenza A Virus in Wild Birds

Björn Olsen, et al.

Science 312, 384 (2006);
DOI: 10.1126/science.1122438

This copy is for your personal, non-commercial use only.

If you wish to distribute this article to others, you can order high-quality copies for your
colleagues, clients, or customers by clicking here.

Permission to republish or repurpose articles or portions of articles can be obtained by

following the guidelines here.

The following resources related to this article are available online at www.sciencemag.org
(this information is current as of October 18, 2010 ):

Updated information and services, including high-resolution figures, can be found in the online
version of this article at:

Downloaded from www.sciencemag.org on October 18, 2010

http://www.sciencemag.org/cgi/content/full/312/5772/384

Supporting Online Material can be found at:

http://www.sciencemag.org/cgi/content/full/312/5772/384/DC1
A list of selected additional articles on the Science Web sites related to this article can be
found at:
http://www.sciencemag.org/cgi/content/full/312/5772/384#related-content
This article cites 31 articles, 16 of which can be accessed for free:
http://www.sciencemag.org/cgi/content/full/312/5772/384#otherarticles

This article has been cited by 248 article(s) on the ISI Web of Science.

This article has been cited by 62 articles hosted by HighWire Press; see:
http://www.sciencemag.org/cgi/content/full/312/5772/384#otherarticles

This article appears in the following subject collections:

Epidemiology
http://www.sciencemag.org/cgi/collection/epidemiology

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the
American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright
2006 by the American Association for the Advancement of Science; all rights reserved. The title Science is a
registered trademark of AAAS.
 INFLUENZA

distance migrations (6), thereby potentially

REVIEW distributing LPAI viruses between countries or
even continents. Birds breeding in one geographic
Global Patterns of Influenza A region often follow similar migratory flyways,
e.g., the East Asian–Australian flyway from
Virus in Wild Birds eastern Siberia south to eastern Asia and Australia
(Fig. 1A). However, the major flyways are
simplifications, and there are numerous excep-
Björn Olsen,1,2 Vincent J. Munster,3 Anders Wallensten,4,5 Jonas Waldenström,6
tions where populations behave differently from
Albert D. M. E. Osterhaus,3 Ron A. M. Fouchier3*
the common patterns (6, 8). Within the large
The outbreak of highly pathogenic avian influenza of the H5N1 subtype in Asia, which has continents and along the major flyways, migra-
subsequently spread to Russia, the Middle East, Europe, and Africa, has put increased focus on tion connects many bird populations in time and
the role of wild birds in the persistence of influenza viruses. The ecology, epidemiology, genetics, space, either at common breeding areas, during
and evolution of pathogens cannot be fully understood without taking into account the ecology migration, or at shared nonbreeding areas (Fig.
of their hosts. Here, we review our current knowledge on global patterns of influenza virus 1). As a result, virus-infected birds can transmit
infections in wild birds, discuss these patterns in the context of host ecology and in particular their pathogens to other populations that subse-

Downloaded from www.sciencemag.org on October 18, 2010

birds’ behavior, and identify some important gaps in our current knowledge. quently may bring the viruses to new areas.
It is important to realize that the transmission
nfluenza A viruses have been isolated from HPAI isolates have been obtained primarily of the viruses and their geographical spread is

I many species, including humans, pigs, horses,

mink, felids, marine mammals, and a wide
range of domestic birds, but wildfowl and
from commercially raised poultry (3).
In the past decade, HPAI outbreaks have
occurred frequently, caused by influenza viruses
dependent on the ecology of the migrating hosts.
For instance, migrating birds rarely fly the full
distance between breeding and nonbreeding areas
shorebirds are thought to form the virus res- of subtype H5N1 in Asia, Russia, the Middle without stopping over and ‘‘refueling’’ along the
ervoir in nature. The influenza A virus genome East, Europe, and Africa (ongoing since 1997); way. Rather, birds make frequent stopovers dur-
consists of eight segments of negative-stranded H5N2 in Mexico (1994), Italy (1997), and ing migration and spend more time eating and
RNA, which code for 11 proteins. Influenza Texas (2004); H7N1 in Italy (1999); H7N3 in preparing for migration than actively performing
viruses are classified on the basis of two of these Australia (1994), Pakistan (1994), Chile (2002), flights (9). Many species aggregate at favorable
proteins expressed on the surface of virus and Canada (2003); H7N4 in Australia (1997); stopover or wintering sites, resulting in high local
particles; the hemagglutinin (HA) and neur- and H7N7 in the Netherlands (2003) (3, 4). densities. Such sites may be important for trans-
aminidase (NA) glycoproteins (1). In wild birds mission of LPAI viruses between wild and cap-
and poultry throughout the world, influenza Migratory Birds as a Natural Reservoir tive birds and between different species.
viruses representing 16 HA and 9 NA antigen- of LPAI Viruses
ic subtypes have been detected (2), which can LPAI viruses have been isolated from at least 105 Influenza Viruses in Ducks
be found in numerous combinations (also wild bird species of 26 different families (Table 1) Extensive surveillance studies of wild ducks in
called subtypes, e.g., H1N1, H16N3). (5). All influenza virus subtypes and most HA/NA the Northern Hemisphere have revealed high
The HA protein is initially synthesized as a combinations have been detected in the bird LPAI virus prevalence primarily in juvenile—
single polypeptide precursor (HA0), which is reservoir and poultry, whereas relatively few have presumably immunologically naı̈ve—birds with a
cleaved into HA1 and HA2 subunits by proteases. been detected in other species. Although many peak in early fall before southbound migration. In
The mature protein mediates binding of the virus wild bird species may harbor influenza viruses, North America, the prevalence falls from È60%
to host cells, followed by fusion with endosomal birds of wetlands and aquatic environments such in ducks sampled at marshalling sites close to the
membranes (1). Influenza viruses of subtypes H5 as the Anseriformes (particularly ducks, geese, Canadian breeding areas in early fall, to 0.4 to 2%
and H7, but not other HA subtypes, may become and swans) and Charadriiformes (particularly at the wintering grounds in the southern U.S.A.,
highly pathogenic after introduction into poultry gulls, terns, and waders) constitute the major nat- and È0.25% on the ducks’ return to the breeding
and can cause outbreaks of highly pathogenic ural LPAI virus reservoir (1). Anseriformes and grounds in spring. Similar patterns have been
avian influenza (HPAI, formerly termed Bfowl Charadriiformes are distributed globally, except observed in Northern Europe, but influenza virus
plague[). The switch from a low pathogenic for the most arid regions of the world (6). detection during spring migration can be signif-
avian influenza (LPAI) virus phenotype, common In birds, LPAI viruses preferentially infect cells icantly higher, up to 6.5%. Surveillance of the
in wild birds and poultry, to the HPAI virus lining the intestinal tract and are excreted in high nesting grounds of ducks in Siberia before winter
phenotype is achieved by the introduction of concentrations in their feces. It has been shown that migration revealed the presence of influenza
basic amino acid residues into the HA0 cleavage influenza viruses remain infectious in lake water viruses in up to 8% of birds (10).
site, which facilitates systemic virus replication. up to 4 days at 22-C and more than 30 days at 0-C Such year-round prevalence raises the pos-
(7), and the relatively high virus prevalence in sibility that LPAI virus can persist in ducks
1
Department of Infectious Diseases, Umeå University, SE- birds living in aquatic environments may be due alone. This hypothesis complements earlier
90187 Umeå, Sweden. 2Section for Zoonotic Ecology and in part to efficient transmission through the fecal- ones, in which additional host species or
Epidemiology, Department of Biology and Environmental
oral route via surface waters (1, 7). preservation of infectious influenza viruses in
Science, University of Kalmar, SE-39182 Kalmar, Sweden.
3
Department of Virology, Erasmus Medical Center, Rotter- Migration is a common strategy for birds frozen lakes over the winter play a role in the
dam, Netherlands. 4Smedby Health Center, Kalmar County occupying seasonal habitats and may range from perpetuation of avian influenza viruses (1, 7).
Council, SE-39471 Kalmar, Sweden. 5Division of Virology, short local movements to intercontinental migra- All HA and NA subtypes, with the exception
Department of Molecular and Clinical Medicine (IMK) tions. Migratory birds can carry pathogens, par- of H13 to H16, circulate in wild ducks in North
Faculty of Health Sciences, Linköping University, SE-581
85 Linköping, Sweden. 6Department of Animal Ecology, ticularly those that do not significantly affect the America and Northern Europe. In a 26-year
Lund University, SE-22362 Lund, Sweden. birds’ health status and consequently interfere longitudinal study performed in Canada, influenza
*To whom correspondence should be addressed. E-mail: with migration. Many Anseriformes and Charadrii- viruses of subtypes H3, H4, and H6 were isolated
r.fouchier@erasmusmc.nl formes are known to perform regular long- from ducks most frequently; H1, H2, H7, H10,

384 21 APRIL 2006 VOL 312 SCIENCE www.sciencemag.org

 SPECIALSECTION
and H11 less frequently; and H5, H8, H9, and H12 insights in the role of spatial factors, herd im- sively studied species, have been found to be in-
only sporadically. Although in other North munity, and population age-structure on epidemi- fected with influenza viruses more frequently
American and European studies, influenza viruses ology (12). Cycling of influenza virus in wild than other birds, including diving ducks (Table 1)
of subtypes H3, H4, and H6 were also detected birds could provide similar new insights into the (5). Differences in virus prevalence between
frequently, the detection of other virus subtypes ecology of influenza viruses in their natural hosts. ecological guilds of ducks are likely in part
was not significantly different (4, 11). Thus, the Influenza virus surveillance of ducks has been related to behavior. Dabbling ducks feed primar-
prevalence of influenza virus in general, as well performed in Japan since the late 1970s. As in ily on food in surface waters; diving ducks
as the specific distribution of subtypes, may vary other studies, influenza virus prevalence and iso- forage at deeper depths and more often in
between different surveillance studies depending lated subtypes varied between years and locations marine habitats (6). Dabbling ducks display a
on species, time, and place. (5). The prevalence of influenza virus in wild propensity for abmigration, the switching of
In the Canadian studies, cyclic patterns of in- birds elsewhere in Asia is largely unknown, but breeding grounds between years, which is in
fluenza virus subtypes were reported: Peaks in several studies have been conducted in live bird part due to mate choice (6). This behavior
virus isolation of an HA subtype were followed 1 markets, where most HA and NA subtypes were could provide an opportunity for influenza
to 2 years later by reduced rates of isolation of this found in poultry (1, 13). It is plausible that the viruses to be transmitted between different host
subtype. This observation awaits confirmation in circulation of the LPAI virus subtypes in poultry subpopulations. LPAI virus infection generally
other surveillance studies but is of particular inter- at least partially reflects that in wild birds, but no causes no major clinical signs in dabbling ducks,

Downloaded from www.sciencemag.org on October 18, 2010

est in relation to findings for other infectious dis- direct connection has yet been established. and experimental infections indicate that animals
eases: Cyclic patterns described for measles and Dabbling ducks of the Anas genus, with only produce a transient, low-level humoral
whooping cough in humans have provided new Mallards (Anas platyrhynchos) as the most exten- immune response, which may be sufficient to
provide partial protection against reinfection
Table 1. Prevalence of influenza A virus in wild birds. Influenza virus prevalence in specific species is given with viruses of the same subtype but is unlikely
only if tests on 9500 birds have been reported; lower numbers in individual species are included in the total. to confer protection against heterologous rein-
See (5) for additional comments and original data. Of the 36 species of ducks, 28,955 were dabbling fections (14). Different influenza virus subtypes
ducks and 1011 were diving ducks, with influenza virus prevalence of 10.1 and 1.6%, respectively. can also infect ducks concomitantly, creating
the opportunity for genetic mixing (15).
Positive Little is known about the prevalence of
Family Species Sampled (n) (%) influenza viruses in wild ducks in the Southern
Hemisphere or potential transmission between
Ducks 36 species 34,503 3275 9.5
the hemispheres. There is little connectivity
Mallard (Anas platyrhynchos) 15,250 1965 12.9
between northern and southern Anatidae spe-
Northern Pintail (Anas acuta) 3,036 340 11.2
cies, and most species stay year round within
Blue-winged Teal (Anas discors) 1,914 220 11.5
each breeding continent. The Blue-winged Teal
Common Teal (Anas crecca) 1,314 52 4.0
(Anas discors) is one of the few North Amer-
Eurasian Wigeon (Anas penelope) 1,023 8 0.8
ican species that has a winter distribution that
Wood Duck (Aix sponsa) 926 20 2.2
includes South America (Fig. 1C) (6). There
Common Shelduck (Tadorna tadorna) 881 57 6.5
are several other duck species that could serve
American Black Duck (Anas rubripes) 717 130 18.1
as hosts for influenza virus in South America
Green-winged Teal (Anas carolinensis) 707 28 4.0
(6), but surveillance data are not available.
Gadwall (Anas strepera) 687 10 1.5
Similarly, only 6 of 39 Anatidae species breed-
Spot-billed Duck (Anas poecilorhyncha) 574 21 3.7
ing in Eurasia winter with at least part of the
Geese 8 species 4,806 47 1.0
population south of the Sahara desert in Africa,
Canada Goose (Branta canadensis) 2,273 19 0.8
e.g., the Garganey (Anas querquedula) (Fig.
Greylag Goose (Anser anser) 977 11 1.1
1C) and the Northern Pintail (Anas acuta), each
White-fronted Goose (Anser albifrons) 596 13 2.2
have African winter populations in excess of
Swans 3 species 5,009 94 1.9
one million birds (16). As in South America,
Tundra Swan (Cygnus columbianus) 2,137 60 2.8
none of the 22 Anatidae species that breed in
Mute Swan (Cygnus olor) 1,597 20 1.3
sub-Saharan Africa spend the nonbreeding
Whooping Swan (Cygnus cygnus) 930 14 1.5
season outside the continent. However, there
Gulls 9 species 14,505 199 1.4
are several species with large, widespread pop-
Ring-billed gull (Larus delawarensis) 6,966 136 2.0
ulations in Africa (16), and some migrate within
Black-tailed Gull (Larus crassirostris) 1,726 17 1.0
Africa (17). Potential areas for mixing of
Black-headed Gull (Larus ridibundus) 770 17 2.2
Eurasian and African ducks are in West Africa,
Herring Gull (Larus argentatus) 768 11 1.4
near the Senegal and Niger Rivers, the flood-
Mew Gull (Larus canus) 595 0 0.0
plains of the Niger River in Nigeria and Mali,
Terns 9 species 2,521 24 0.9
and Lake Chad (16), and influenza viruses in
Common Tern (Sterna hirundo) 961 16 1.7
African Anatidae populations may thus be linked
Waders 10 species 2,637 21 0.8
to Eurasia through migrating species. Anatidae of
Rails 3 species 1,962 27 1.4
Oceania are mainly resident and do not perform
Eurasian Coot (Fulica atra) 1,861 23 1.2
regular seasonal migrations (6).
Petrels 5 species 1,416 4 0.3
Wedge-tailed Shearwater (Puffinus pacificus) 794 4 0.5 Influenza Viruses in Gulls and Terns
Cormorants 1 species 4,500 18 0.4
The first recorded isolation of influenza virus from
Great Cormorant (Phalacrocorax carbo) 4,500 18 0.4
wild birds was from a Common Tern (Sterna

www.sciencemag.org SCIENCE VOL 312 21 APRIL 2006 385

 INFLUENZA

hirundo) in 1961. This HPAI perpetuation of certain virus

H5N3 virus was responsible A East subtypes. Influenza viruses of
Black Sea–
Atlantic
for an outbreak in South Af- Mississippi Mediterranean subtypes H1 to H12 have been
flyway East Asia
rica where at least 1300 of Americas flyway isolated in birds migrating
Australian
these birds died (3). The most flyway flyway through the eastern U.S.A.,
frequently detected LPAI vi- Atlantic Central Asia with a high prevalence of
rus subtype in gulls is H13, a Americas flyway certain HA subtypes (H1, H2,
subtype rarely found in other flyway H5, H7, H9 to H12) and a
birds. Recently, a ‘‘novel’’ larger variety of HA/NA com-
virus subtype (H16), related binations as compared with
to H13, was described in ducks in Canada, suggesting
East Africa
Black-headed Gulls (Larus Pacific
West Asia
that waders maintain a wider
ridibundus) in Sweden. The Americas spectrum of viruses. More-
flyway flyway
genes of H13 and H16 viruses over, the seasonal prevalence
are genetically distinct from of influenza viruses in waders
those of influenza viruses seems to be reversed as com-

Downloaded from www.sciencemag.org on October 18, 2010

from other hosts, which sug- pared with ducks, with higher
gests they have been geneti- virus prevalence (È14%) dur-
cally isolated for sufficient B ing spring migration (19). This
time to allow genetic dif- has led to the hypothesis that
ferentiation (2). This concurs different families of wetland
with the observation that gull birds are involved in perpetu-
influenza viruses do not read- ation of LPAI virus and sug-
ily infect ducks when they gests a role for waders, which
are inoculated experimental- may carry the virus north to
ly (1). Although other influ- the duck breeding grounds in
enza virus subtypes are also spring. Recent genetic analy-
occasionally detected in terns ses have not revealed striking
and gulls (Table 1) (5), it is differences between influenza
plausible that the viruses that viruses from ducks and wad-
are genetically indistinguish- ers in the Americas, suggest-
able from viruses of other avian ing that these viral gene pools
hosts are most likely not en- are not separated (20, 21).
demic in gulls and terns. Although the wader-duck link
Influenza viruses can be may be a plausible scenario
detected in a small proportion based on the North American
of gulls, with the highest virus C data, studies in waders in
prevalence reported in late Northern Europe have failed
summer and early fall. Most to produce similar results.
gull species breed in colonies Nevertheless, many wader
(6), with adults and juveniles species of the Northern Hem-
crowded in a small space, isphere are long-distance inter-
creating good opportunities continental migrants (8) and
for virus spread. This situation may, therefore, have the po-
contrasts with that in dabbling tential to distribute influenza
ducks that do not breed in viruses around the globe.
dense colonies (6), and epi-
zootics could be more easily Influenza Viruses in Other
initiated when birds congre- Wild Birds
gate in large numbers during LPAI viruses can be found in
molt, migration, or wintering. numerous other bird species
(Table 1) (5), but it is unclear
Influenza Viruses in Waders in which of these species
Fig. 1. Migratory flyways of wild bird populations. A world map with the main general migratory
Waders in the Charadriidae influenza viruses are endemic
flyways of wild bird populations is shown (adapted from information collected and analyzed by
and Scolopacidae families are Wetlands International). (A) Black dots indicate the locations of historical and current influenza and in which the virus is a
adapted to either marine or virus surveillance sites from which data have been used in this manuscript. These global temporary pathogen. Species
freshwater wetland areas and migration flyways are simplifications, and there are situations where populations behave in which influenza viruses are
often live side-by-side with differently from the common patterns. Migration patterns of Mallard (Anas platyrhynchos) (B) endemic share the same hab-
ducks (18). Long-term influ- and Garganey (Anas querquedula) in Eurasia and Africa and Blue-winged Teal (Anas discors) in itat at least part of the year
enza virus surveillance studies the Americas (C) (right and left parts of the map, respectively) are provided. Yellow color with other species in which in-
are still sparse, but data from indicates breeding areas in which species are absent during winter, green indicates areas in fluenza viruses are frequently
North America suggest a dis- which species are present around the year, and blue indicates areas in which species are only detected, including geese,
tinct role of these birds in the present in winter and do not breed. Arrows indicate the seasonal migration patterns. swans, rails, petrels, and cor-

386 21 APRIL 2006 VOL 312 SCIENCE www.sciencemag.org

 SPECIALSECTION
morants. In these birds and others (5), influenza but allows occasional spillover of gene segments genetic reassortment, i.e., the mixing of genes
virus prevalence seems to be lower than in from one gene pool to the other. from two or more influenza viruses. A recent
dabbling ducks (Table 1), but it should be noted Within each genetic lineage, multiple sub- study of 35 influenza virus isolates obtained from
that studies on these species are limited, and it is lineages of viral genes cocirculate, but there appear ducks in Canada indicates that genetic ‘‘sub-
possible that peak prevalence has been missed to be no consistent temporal or spatial correlations. lineages’’ do not persist, but frequently reassort with
because of its seasonal nature or location. Moreover, genetic data from duck and shorebird other viruses (27). Influenza viruses of a particular
As for ducks, gulls, and waders, their behavior influenza virus isolates from the Americas suggest subtype do not necessarily have the same genetic
and ecology may be an important determinant of an active interplay between these host species make-up, even within a single year or a single host
their role as host species. For species. The high prevalence of in-
instance, geese are mainly herbivo- fluenza virus in some wild bird spe-
rous and often congregate in large cies and the sporadic detection of
flocks for grazing in pastures and concomitant infections in single birds
agricultural fields, especially during (15) support the notion that reassort-
the nonbreeding season. Such flocks ment may occur in nature. Gaining
may reach tens of thousands of birds information on the actual frequency
in optimal areas and often contain of reassortment in the wild bird

Downloaded from www.sciencemag.org on October 18, 2010

several different species. Colonial reservoir and the impact of these
breeding occurs in some goose events on LPAI virus evolution will
species, but most are solitary nesters be of considerable interest.
or nest in loose groups with little
interaction between pairs. Given that HPAI H5N1 Viruses in Wild Birds
wild geese and ducks are the ances- In 1997, an HPAI outbreak caused
tors of today’s domestic goose and by H5N1 influenza virus occurred
duck species and that these domestic in chicken farms and the live bird
animals in parts of the world are markets of Hong Kong, which also
frequently kept alongside chickens, resulted in the first reported case of
wild geese and ducks may form the human influenza and fatality attrib-
bridge for influenza viruses between utable directly to avian influenza
wild and domestic birds. virus (28). The H5N1 HPAI virus
reappeared in 2002 in waterfowl at
Genetic Variation of Influenza two parks in Hong Kong and was
Viruses in Wild Birds also detected in other captive and
Evolution of avian influenza viruses American Swine wild birds (29). It resurfaced again
in their natural hosts is slow, but not in 2003 and has devastated the
Human
negligible. Avian influenza viruses poultry industry in large parts of
can be divided into two lineages, American Gull Southeast Asia since 2004. In 2005,
Eurasian and American (Fig. 2), American Avian the virus was isolated during an
probably as a result of long-term eco- outbreak among migratory birds in
logical and geographical separation Equine Qinghai Lake, China, affecting
of hosts. However, the avifauna of Eurasian Swine large numbers of wild birds (30).
North America and Eurasia are not Eurasian Gull This single epizootic caused an es-
completely separated; some ducks timated 10% decrease of the global
and shorebirds cross the Bering Strait Eurasian Avian population of Bar-headed Geese
during migration or have breeding H5N1 HPAI (Anser indicus), highlighting the
ranges that include both the Russian potential devastating effects on
Fig. 2. Phylogenetic tree for the matrix gene of influenza A viruses from a
Far East and northwestern North vulnerable wildlife. Subsequently,
variety of hosts. Nucleotide sequences were selected from public databases and
America (6). The majority of tundra the virus has appeared across Asia,
aligned, after which a maximum likelihood tree was generated using influenza
shorebirds from the Russian Far East virus A/Equine/Prague/57 (H7N7) as outgroup. Sequences were selected from Europe and the Middle East, and in
winter in Southeast Asia and Austra- each host to reflect the longest possible time frame and variation in locations of several African countries. Wild
lia, but some species winter along the virus isolation. The avian influenza viruses are divided in an American lineage bird deaths have been reported in
west coast of the Americas (22). The (pink) and a Eurasian lineage (yellow), and there are no clear patterns of host, several of these countries, in Eu-
overlap in distribution of ducks is not temporal, or spatial correlation within these lineages. In contrast, the human rope, particularly affecting Mute
as profound as that of shorebirds, but influenza A virus lineage (light blue), the Eurasian swine lineage (purple), and the Swans (Cygnus olor) and Whoop-
a few species (e.g., Northern Pintail, HPAI H5N1 lineage (orange) display clear temporal patterns of virus evolution. er Swans (Cygnus cygnus), but
Anas acuta) are common in both mortality has also been recorded
North America and Eurasia (6) and could also (20, 21). Although certain HA subtypes are reported in other waterfowl species, and occasionally in
provide an intercontinental bridge for influenza to be more prevalent in either shorebirds or ducks raptors, gulls, and herons. So far, the HPAI
virus. Indeed, influenza viruses carrying a mix of in North America, this also does not seem to have H5N1 strain that originated in poultry in South-
genes from the American and Eurasian lineages resulted in differences in the genetic composition east Asia has caused mortality in 960 wild bird
have been isolated, indicating that allopatric spe- of influenza viruses obtained from these two species (29–31). In addition, during the dev-
ciation is only partial (23–25). The partial ecological reservoirs (19, 26). astating outbreaks in poultry, the H5N1 virus
isolation of influenza virus hosts seems sufficient to The segmented nature of the influenza virus was transmitted to 175 humans, leading to 95
facilitate divergent evolution of separate gene pools, genome enables evolution by a process known as deaths (as of 6 March 2006), and has also

www.sciencemag.org SCIENCE VOL 312 21 APRIL 2006 387

 INFLUENZA

been isolated from pigs, cats, tigers, and of LPAI viruses in wild birds, it will be crucial vides a unique opportunity to increase our under-
leopards. to integrate virus and host ecology with long- standing not only of HPAI epidemiology but also
It is most likely that the H5N1 virus has term surveillance studies to provide more insight of the ecology of LPAI viruses in their natural
circulated continuously in domestic birds in on the year-round perpetuation of influenza hosts, at the same time and for the same cost.
Southeast Asia since 1997 and, as a consequence, viruses in wild birds. Possible intercontinental
has evolved substantially (Fig. 2). Surveillance contacts among ducks and shorebirds in areas References and Notes
studies in Mainland China from 1999 onward in- where migrating birds from the northern and 1. R. G. Webster, W. J. Bean, O. T. Gorman, T. M. Chambers,
Y. Kawaoka, Microbiol. Rev. 56, 152 (1992).
dicated that H5N1 viruses have become endemic southern latitudes mix are of particular interest.
2. R. A. M. Fouchier et al., J. Virol. 79, 2814 (2005).
in domestic birds in the region and that multiple Can influenza viruses be perpetuated in ducks 3. D. J. Alexander, Vet. Microbiol. 74, 3 (2000).
genetic lineages of the virus are cocirculating alone, or does the interface between ducks and 4. V. J. Munster et al., Emerg. Infect. Dis. 11, 1545 (2005).
(32, 33). Poultry trade and mechanical move- shorebirds, as seems to occur in North America 5. Table S1 and references are available as supporting
material on Science Online.
ment of infected materials are likely modes for (19), also occur on other continents? With high-
6. J. Del Hoyo, A. Elliot, J. Sargatal, Eds., Handbook of the
spreading HPAI in general (3). For the H5N1 throughput sequencing technology, it should be Birds of the World (Lynx Edicions, Barcelona, 1996),
virus, it is without doubt that domestic waterfowl, possible to gain more insight into the genetic vols. 1 and 3.
specific farming practices, and agroecological variability and evolution of LPAI viruses in wild 7. R. G. Webster, M. Yakhno, V. S. Hinshaw, W. J. Bean, K. G.
environments played a key role in the occurrence, birds and to integrate this information with Murti, Virology 84, 268 (1978).
8. J. van de Kam, B. Ens, T. Piersma, L. Zwarts, Shorebirds:

Downloaded from www.sciencemag.org on October 18, 2010

maintenance, and spread of HPAI for many epidemiology and virus-host ecology. An Illustrated Behavioural Ecology (KNNV Publishers,
affected countries (34, 35). Although numerous The recent H5N1 outbreaks in Eurasia have Utrecht, Netherlands, 2004).
wild birds have also become infected, it has been identified additional gaps in our knowledge of 9. T. Alerstam, A. Lindstrom, in Bird Migration: Physiology
much debated whether they play an active role in avian influenza viruses in wild birds in general. and Ecophysiology, E. Gwinner, Ed. (Springer-Verlag,
Berlin, 1990), pp. 331–351.
the geographic spread of the disease. It has been It should be realized that our knowledge of 10. K. Okazaki et al., Arch. Virol. 145, 885 (2000).
argued that infected birds would be too severely LPAI viruses in wild birds cannot simply be 11. B. A. Hanson, D. E. Stallknecht, D. E. Swayne, L. A. Lewis,
affected to continue migration and thus unlikely extrapolated to HPAI viruses; for instance, the D. A. Senne, Avian Dis. 47, 867 (2003).
to spread the H5N1 virus. Although this may be most important host species or routes of 12. P. Rohani, D. J. Earn, B. T. Grenfell, Science 286, 968 (1999).
13. D. C. Nguyen et al., J. Virol. 79, 4201 (2005).
true for some wild birds, it has been shown that, in transmission may be quite different (Table 1)
14. H. Kida, R. Yanagawa, Y. Matsuoka, Infect. Immun. 30,
experimental infections, several bird species sur- (29–31, 38). It is clear that influenza virus sur- 547 (1980).
vive infection and shed the H5N1 virus without veillance of wild birds could provide ‘‘early 15. G. B. Sharp et al., J. Virol. 71, 6128 (1997).
apparent disease signs (31, 32, 36). In addition, warning’’ signals for the introduction of HPAI 16. D. A. Scott, P. M. Rose, Atlas of Anatidae Populations in
Africa and Western Eurasia (Publ. no. 41, Wetlands
many wild birds may be partially immune owing H5N1 virus in new regions and may provide
International, Wageningen, Netherlands, 1996).
to previous exposures to LPAI influenza viruses, access to strains for characterization. For 17. L. H. Brown, E. K. Urban, K. Newman, The Birds of Africa
as has been shown for chickens (37). Finally, proper risk assessment studies, however, we (Academic Press, London, 1982), vol. 1.
recent studies suggest that HPAI viruses may also need a better understanding of the in- 18. T. Piersma, Oikos 80, 623 (1997).
become less pathogenic to ducks infected exper- terface between wild and domestic birds, the 19. S. Krauss et al., Vector Borne Zoonotic Dis. 4, 177 (2004).
20. L. Widjaja, S. L. Krauss, R. J. Webby, T. Xie, R. G. Webster,
imentally, while retaining high pathogenicity for possible transmission of influenza viruses be- J. Virol. 78, 8771 (2004).
chickens (32, 36, 38). The present situation in tween these populations, bird behavior, age- 21. E. Spackman et al., Virus Res. 114, 89 (2005).
Europe, where infected wild birds have been structures of populations, and detailed migration 22. R. E. Gill, R. W. Butler, P. S. Tomkovich, T. Mundkur, C. M.
found in several countries that have not reported routes. We further need better understanding of Handel, Trans. North Am. Wildlife Nat. Resources Conf.
59, 63 (1994).
outbreaks among poultry, suggests that wild the transmission and pathogenesis of H5N1 23. A. Wallensten et al., Arch. Virol. 150, 1685 (2005).
birds can indeed carry the virus to previously virus in wild birds, as well as identification of 24. N. V. Makarova, N. V. Kaverin, S. Krauss, D. Senne, R. G.
unaffected areas. Although swan deaths have virus-permissive host species and their relative Webster, J. Gen. Virol. 80, 3167 (1999).
been the first indicator for the presence of the likelihood to develop disease, patterns of virus 25. J. H. Liu et al., Virus Genes 29, 81 (2004).
26. J. C. Obenauer et al., Science 311, 1576 (2006).
H5N1 virus in several European countries, this secretion, and temporal and spatial variations
27. T. F. Hatchette et al., J. Gen. Virol. 85, 2327 (2004).
does not necessarily imply a role as predom- in virus prevalence. 28. J. C. de Jong, E. C. Claas, A. D. Osterhaus, R. G. Webster,
inant vectors; they could merely have func- With our current limited knowledge on W. L. Lim, Nature 389, 554 (1997).
tioned as sentinel birds infected via other HPAI in wild birds, there is no solid basis for 29. T. M. Ellis et al., Avian Pathol. 33, 492 (2004).
migrating bird species. including wild birds in control strategies be- 30. J. Liu et al., Science 309, 1206 (2005).
31. K. M. Sturm-Ramirez et al., J. Virol. 78, 4892 (2004).
yond the physical separation of poultry from 32. H. Chen et al., Proc. Natl. Acad. Sci. U.S.A. (2006).
Prospects wild birds. Even in areas with significant 33. K. S. Li et al., Nature 430, 209 (2004).
Despite the relatively intense surveillance studies outbreaks in poultry, virus prevalence in wild 34. V. Martin et al., Dev. Biol. 124, 23 (2006).
that have been performed for many years in North birds is low (32), and the role of these wild 35. M. Gilbert et al., Emerg. Infect. Dis. 12, 227 (2006).
36. D. J. Hulse-Post et al., Proc. Natl. Acad. Sci. U.S.A. 102,
America and Eurasia, our understanding of the birds in spreading the disease is unclear. It is 10682 (2005).
global distribution of LPAI viruses in wild bird clear that the H5N1 problem originated from 37. S. H. Seo, M. Peiris, R. G. Webster, J. Virol. 76, 4886 (2002).
populations is still limited. Serological evidence outbreaks in poultry and that the outbreaks 38. K. M. Sturm-Ramirez et al., J. Virol. 79, 11269 (2005).
indicates that influenza viruses occasionally cir- and their geographical spread probably cannot 39. F. J. Austin, R. G. Webster, J. Wildl. Dis. 29, 568 (1993).
40. We apologize to the scientists whose original contributions
culate in Antarctica (39), and it is reasonable to be stopped without implementation of proper have not been cited here due to space restrictions. We thank
assume that influenza viruses are distributed control measures in the global poultry indus- G. Rimmelzwaan, D. Smith, T. Piersma, J. de Jong, and
globally, wherever competent host species are try. However, there is at present no scientific E. de Wit for fruitful discussions and helpful comments.
present. It is possible that some subtypes are rare basis for culling wild birds to control the
or not detected annually in current surveillance outbreaks and their spread, and this is further Supporting Online Material
www.sciencemag.org/cgi/content/full/312/5772/384/DC1
studies. Simply because of the limitations of our highly undesirable from a conservationist Table S1
studies, we are currently biased toward species perspective. References and Notes
that are easy to sample during migration or win- The current increased interest in influenza
tering. Second, to understand the global patterns virus surveillance in wild and domestic birds pro- 10.1126/science.1122438

388 21 APRIL 2006 VOL 312 SCIENCE www.sciencemag.org