Introduction

The highly pathogenic avian influenza (HPAI) panzootic caused by H5 viruses of clade 2.3.4.4b is unprecedented in scale and impact. Since 2020, HPAI H5 viruses have spread to numerous countries worldwide, reaching over 76 countries across Europe, Asia, Africa, the Americas1, and even Antarctica2, causing massive die-offs in wild birds, severe outbreaks in poultry, and increasing spillovers to mammals3,4,5,6. While this shift in viral dynamics is recent, the history of avian influenza viruses extends far into the past.

Historically, avian influenza viruses have been naturally maintained in wild waterbirds, particularly those in the orders Anseriformes and Charadriiformes7, as low pathogenic forms, with HPAI cases reported only sporadically in wild birds8. HPAI outbreaks were mostly confined to poultry, typically arising when low pathogenic forms introduced from wild waterbirds evolved into highly pathogenic forms9. However, the emergence of the A/Goose/Guangdong/1/1996 (Gs/Gd) HPAI H5 lineage in 1996 in China marked a critical turning point. This lineage became established in poultry, and spillovers to wild bird populations were first reported in 200210. From 2005, wild birds cases became more frequent, spreading across Asia, Europe and Africa11,12. Over time, the Gs/Gd HPAI H5 lineage diversified into multiple clades, with clade 2 increasingly detected in wild birds13,14,15. In 2014, the clade 2.3.4.4 spread beyond Asia16, with migratory wild birds playing a key role in spatial dissemination16,17,18.

Since 2020, the spread of HPAI H5 clade 2.3.4.4b has escalated into a global panzootic19,20,21. While circulating in Eurasia and Africa in 2020, HPAI H5 viruses surged in North America in 2021 and reached South America in 202222,23. The 2020–2021 wave was primarily driven by the H5N8 subtype24,25 with a peak in wild bird detections in winter followed by a sharp decline in summer months (Fig. 2A, Supplementary Fig. 1)24. The H5N1 subtype completely dominated in subsequent waves25,26 and persistent viral circulation in summers 2022 and 2023 was observed, differing from the usual HPAI seasonality and affecting a wide spectrum of hosts, notably colonial breeding birds. Globally, reported cases have encompassed over 500 wild bird species spanning 25 orders, with more than half of the species never reported as infected before 202127. The death toll on wild birds has been particularly high, though difficult to quantify. Some species, previously spared, have experienced mass mortality events, with potential long-term consequences. The ongoing viral circulation in wild birds has also fueled large-scale outbreaks in poultry1,28 and contributed to repeated spillovers in wild and domestic mammals3,4,5,6, highlighting the role of wild birds in cross-species transmission. The rising incidence of HPAI H5 viruses in mammals is particularly alarming29, as it raises concerns about viral adaptation to these hosts, posing risks to public health.

Wild birds have emerged as both victims and vectors of HPAI H5 viruses, experiencing unprecedented mortality while driving the virus’s spatial dispersion and seeding outbreaks in poultry and mammals. A better understanding of their role in this global panzootic is therefore essential to study these new viral dynamics and gain broader insights into HPAI H5 epidemiology. In this review, we systematically examine studies on HPAI H5 viruses in wild birds since 2020 to trace the global routes of viral dissemination. We identify key wild bird species involved, examining infection patterns, mortality and asymptomatic cases, as well as transmission pathways linked with their ecological behaviors. Finally, we identify major research gaps that hinder a full understanding of wild birds’ role in HPAI dynamics - gaps that must be addressed to improve effective surveillance and prevention strategies.

Results

Global dissemination routes of HPAI H5 viruses via wild birds

Spatial spread in Eurasia in 2020–2023

In summer 2020, HPAI H5 viruses were detected in North Asia (Russia, Kazakhstan)30,31, with viruses likely originating from North Africa or the Middle East25,32,33 (Fig. 1). North Asia has likely been acting as a key viral hub, due to its role as a major breeding ground for many migratory bird populations and its location at the intersection of multiple flyways33,34. Phylogeographic analyses indicated that by autumn 2020, viruses had disseminated widely from North Asia to Central Asia (China)30,31,33,35,36,37,38, East Asia (Japan and South Korea)33,39, South Asia38 and Europe25,30,31,32,33,34,38,39,40,41,42,43,44 (Fig. 1). Similar dissemination patterns from North Asia to East Asia45,46,47,48 and Europe25,49,50 were reported in autumn 2021 and 2022. Eastward viral movements from Europe to Asia were also documented, with viruses detected in North, Central and East Asia in autumn 2020 found closely related to European ones30,31,33,35,36,37,38,39,51,52,53,54, a pattern repeated in autumn 202136,54,55,56,57 and 202258 (Fig. 1). Frequent bidirectional viral circulation between Central and East Asia in 2020–2023 was also supported by phylogeographic analyses, potentially in both autumn and spring30,31,34,37,38,59. These extensive viral disseminations in Eurasia were closely and timely aligned with wild bird migrations34,38,60. In the Northern Hemisphere, migration is often categorized into autumn movements, when birds travel from breeding to more southerly wintering grounds, and spring movements when they return to breeding grounds. Birds engaging in these journeys while infectious have thus the potential to disseminate viruses along their routes.

Fig. 1: Global dissemination routes of HPAI H5 viruses via wild birds from 2020 to 2023.
figure 1

Red dots represent HPAI H5 cases (H5Nx, H5N8, H5N1) in wild birds, as reported to World Animal Health Information System (WAHIS) of the World Organization for Animal Health. To reduce biases due to reporting inconsistencies across space and time, reports form WAHIS were pooled by months, species and locations. Each map shows an epidemiological season, defined from October 1st to September 30th (extended to the end of 2023 for the 2022–2023 wave). Regions were defined on epidemiological, geographical and ecological criteria. Arrows represent viral dissemination linked to wild bird movements based on the systematic literature review, with shape showing different levels of evidence based on phylogenetic (moderate) or phylogeographic analysis (strong support: Bayes Factor (BF) < 100; decisive support: BF > 100). Links to bird migratory networks were made based on temporality as well as ecological criteria. Black icons indicate major breeding and wintering grounds shared by various species. Grey arrows represent previously observed viral movements that could still be ongoing in the current season.

Full size image

Among wild birds, Anatidae, particularly dabbling ducks, have been considered as primary drivers of viral spread in Eurasia. Movement tracking of infected mallards (Anas platyrhynchos) in China showed it was capable of movements while infected61 and other individuals were found alive while infected across Eurasia, with or without clinical signs62,63,64,65,66,67. Other species found alive while infected in Eurasia included Eurasian wigeons (Mareca penelope)63,64,65,66,67,68, Eurasian teals (Anas crecca)62,66,67,69, spot-billed ducks (Anas poecilorhyncha), northern pintails (Anas acuta), falcated teals (Mareca falcata)62 and northern shovelers (Spatula clypeata)66. In East Asia, phylodynamic analyses showed whooper swans (Cygnus cygnus) could have played an important role in viral introductions30,47, while phylogeny suggests hooded cranes (Grus monacha) could be involved in regional dissemination70. Geese also appeared involved in viral movements in Europe as barnacle (Branta leucopsis), greylag (Anser anser), and pink-footed (Anser brachyrhynchus) geese were alive while infected, with or without clinical signs63,64,65,67.

Following viral introductions from North Asia into Europe, cases in migratory Anatidae surged around October of each year32,40,49,50. Cases were particularly clustered along the coasts of the Wadden and Baltic seas in Germany, the Netherlands, the United Kingdom (UK) and Denmark32,40,49,50, though dependent on surveillance efforts. This region seemed important for further dissemination across Europe in 2020–202225,32,34,39,44,71,72, while extensive viral diffusion was observed with cases reported in Iceland, Svalbard, and Jan Mayen islands for the first time in 202250,72.

Intercontinental spread to North America in 2021–2022

In December 2021, HPAI H5 viruses were detected in Canada22. Phylogeographic analyses revealed multiple introductions from Europe across the Atlantic Ocean in 202125,34,54,59,71,73,74 and 202274,75,76 (Fig. 1), probably occurring in two stages. First, birds migrating from European wintering grounds to northern breeding grounds (Iceland, Greenland) may have carried the virus22,72,75. Transmission could then have occurred to birds breeding in these areas but wintering in North America. Several species could have played a role in this intercontinental spread, at least partially, including Eurasian wigeons, barnacle geese, great skuas (Stercorarius skua) or black-headed gulls (Chroicocephalus ridibundus), that were notably found as vagrants from Europe in Canada22. In parallel, multiple viral introductions occurred on the Pacific coast of North America in 2022, likely originating from East Asia, supported by phylogeographic analyses25,74,77,78,79. Bidirectional viral flow through the Pacific seemed to occur as viruses detected in East Asia in autumn 2022 were closely related to those in North America46. Various species moving between both continents were suggested as possible vectors, although evidence remains limited78.

Following introduction, HPAI H5 viruses rapidly spread across North America (Fig. 1)19,71,73,74,78,79, mostly those introduced on the Atlantic coast, with Anatidae considered primary drivers. Movement tracking of infected mallards and lesser scaups (Aythya affinis) showed their ability to move while infected80,81. Other species found alive while infected include mallards, American green-winged teals (Anas crecca carolinensis), American wigeons (Mareca americana), northern pintails, American black ducks (Anas rubripes), cackling geese (Branta hutchinsii) and snow geese (Anser caerulescens)73,79,82,83. A spatial transmission modeling study also suggested that other species, like pelicans, may have contributed to viral spread84.

Dissemination to Central and South America in 2022–2023

From late 2022, HPAI H5 viruses reached Central and South America via multiple independent viral introductions from North America to Colombia, Peru, Ecuador and Venezuela, as shown by phylogeographic analyses23,85,86,87,88 (Fig. 1). These introductions coincided with autumn migrations to South American wintering grounds89, although the specific species involved remain unclear. The introduction to Peru was particularly impactful, leading to regional spread within Peru and Chile and mass mortality in wild birds and marine mammals23,90,91,92. Subsequent viral dissemination occurred in 2023 to Uruguay87,93,94, Argentina6,95,96 and Brazil90,97,98, as shown by phylogeographic analyses or phylogeny. By late 2023, HPAI H5 viruses have spread from South America to the subantarctic region2,99,100. Brown skuas (Stercorarius antarcticus) and giant petrels (Macronectes spp.) were among the first affected species and could have contributed to the introduction in this region99,100. Other seabird species with migratory connectivity to the South American coast may also have played a role101.

Circulation in Africa in 2020–2023

HPAI H5 viruses were also reported in Africa in 2020–2023, although available data remain limited. Initial viral introductions were detected in West Africa, notably in Senegal and Nigeria in late 2020, likely originating from Europe via autumn migration, as shown by phylogeographic analyses (Fig. 1)25,44,54,102,103. In Nigeria, white storks (Ciconia ciconia) and ruffs (Calidris pugnax), both migrants from Europe to Africa, were found healthy while infected, suggesting their possible role in viral introduction103. Viruses detected in wild birds in Egypt in 2021 were also closely related with those circulating in Europe104.

Evidence of viral dissemination within Africa has been observed, as viruses detected in southern Africa and Nigeria in 2021–2022 were genetically linked to those from Senegal in late 2020 and early 2021102,103. In Nigeria, intra-African migratory species such as African jacanas (Actophilornis africanus), and resident birds like white-faced whistling ducks (Dendrocygna viduata), spur-winged geese (Plectropterus gambensis) and square-tailed nightjars (Caprimulgus fossii) were found healthy while infected, indicating potential for local viral dissemination103. Backflow of viruses from Africa to Europe also likely occurred via spring migration in 2020–2023, as shown by phylogeographic analyses, although less documented (Fig. 1)54,59,105.

Impacts of HPAI H5 viruses on wild bird populations

Unprecedented mortalities in colonial breeding birds

HPAI H5 viruses have caused unprecedented mortality in colonial breeding birds, particularly from summer 2022 with many species reported infected for the first time27. Quantifying the full impact is challenging, as a single reported case could represent thousands of infected birds (Fig. 2A, Table 1), and many events likely went unreported. Among seabirds, northern gannets (Morus bassanus) were among the most severely affected species65, with approximately 75% of North Atlantic colonies impacted in summer 2022106 (Fig. 2F, G, Table 1). In Europe, several colonies were affected including Scotland, with a 70% reduction in colony size and a 75% decline in breeding success106, and France, with adult mortality exceeding 58% and 67% for chicks107. In North America, 11.5% of the population was lost with reproductive success dropping to 17% in one colony108,109,110. Russian colonies also underwent a 40% decline in inhabited nests111. Despite these severe losses, recovery signs have emerged, with birds developing antibodies and iris color changes potentially indicating prior infection106,107,111,112. Other seabird species experienced significant mortality (Table 1), like great skuas, in Great Britain in summer 2021 and 2022, with up to 11% breeding population decline113. Common murres (Uria aalge), Atlantic puffins (Fratercula arctica) and razorbills (Alca torda)108,109,114 were affected in North America, while Peruvian boobies (Sula variegata)85,88,91,115,116,117, Humbolt penguins (Spheniscus humboldti)118, and magnificent (Fregata magnificens) and great frigatebirds (Fregata minor)119 suffered losses in South America. Snowy albatrosses (Diomedea exulans), gentoo penguins (Pygoscelis papua) and brown skuas were affected in the Antarctic region2,99,100, with recoveries after infection observed for this species. In South Africa, African penguins (Spheniscus demersus) were also notably impacted102.

Fig. 2: Impact of HPAI H5 viruses in wild bird populations from 2020 to 2023.
figure 2

Top panel: A Bar plots represent the temporal distribution of HPAI H5 cases (H5Nx, H5N8, H5N1) in wild birds, as reported to World Animal Health Information System (WAHIS) of the World Organization for Animal Health. To reduce biases due to reporting inconsistencies across space and time, reports form WAHIS were pooled by months, species and locations. Seasons were defined from October 1st to September 30th (extended to the end of 2023 for the 2022–2023 wave). Species were categorized into six groups based on taxonomic, epidemiological, and ecological criteria: Seabirds (excluding Laridae), Laridae, Other waterbirds, Anatidae, Raptors, and Other land birds (Supplementary Table 1). The boxplot represents the estimated number of dead birds per case reported in WAHIS for each wild bird group. It shows that quantitative aspects of mortality are not well represented like in mass mortality events, where one case could truly represent thousands. Bottom panel: Pictures of wild birds affected by HPAI H5 viruses: B Juvenile common tern with abandoned eggs in France in summer 2023. C Common tern in France in summer 2023. D Juvenile Sandwich tern in France in summer 2023. E Griffon vulture in France in summer 2022. F, G Northern gannets in France in summer 2022. H Black-legged kittiwakes in Norway in summer 2023. Credits: Bernard Cadiou/Yann Jacob/Thierry David/Pascal Provost/Pierre-Yves Henry.

Full size image
Table 1 Major mortality events caused by HPAI H5 viruses in seabirds, Laridae and other waterbirds
Full size table

Laridae species, including terns and gulls, also experienced devastating outbreaks (Table 1). Sandwich terns (Thalasseus sandvicensis) were particularly affected in summer 2022, with outbreaks in 60% of European colonies, leading to an estimated 17% loss of the total breeding population120 (Fig. 2D, Table 1). Mortality rates exceeded 30% in some German colonies121, while some Dutch colonies saw up to a loss of all breeding adults122. Common terns (Sterna hirundo) also had severe losses in summer 2022 in Germany121, Great Britain123, Canada108 and Ireland in summer 2023124 (Fig. 2B, C, Table 1). Caspian terns (Hydroprogne caspia) lost over 60% of the Lake Michigan population in the United States (US) in summer 202219 and more than half a colony on the Pacific coast in summer 2023125. Other affected tern species included Arctic (Sterna paradisaea), roseate (Sterna dougallii)124, swift (Thalasseus bergii)126 and South American terns (Sterna hirundinacea)6. Among gulls, herring gulls (Larus argentatus) were affected in Europe in 202265 while black-headed gulls became particularly affected from early 202326,105,127,128,129. Other gulls with significant mortality included great black-headed (Ichthyaetus ichthyaetus), Caspian (Larus cachinnans)58, great black-backed (Larus marinus)130 and kelp gulls (Larus dominicanus)115,117 as well as black-legged kittiwakes (Rissa tridactyla)108,109 (Fig. 2H, Table 1). The emergence of a new HPAI H5N1 genotype, which likely emerged from a reassortment with a gull-adapted low pathogenic virus, may have contributed to the high severity of populational impacts25,128. Differential survival between adults and chicks in gulls123,125 and terns120,126, suggests possible acquired immunity, supported by H5 antibodies detection in herring gulls131. In summer 2023, Sandwich terns in the UK showed low mortality, also pointing out possible immune protection124.

Several waterbird species experienced significant population losses in 2020–2023 (Table 1). Dalmatian pelicans (Pelecanus crispus) lost around 60% of a colony in Greece in early 2022132,133 with additional die-offs in Russia69. Peruvian pelicans (Pelecanus thagus) experienced significant mortality along the coasts of Peru and Chile from late 202223,85,86,88,91,115,116,117, while great white pelicans (Pelecanus onocrotalus)44,134,135 and American white pelicans (Pelecanus erythrorhynchos)19,82 were also impacted. Cormorants were also heavily affected, like Cape cormorants (Phalacrocorax capensis) that lost an estimated 30% of their population in Southern Africa102,136. Guanay cormorants (Leucocarbo bougainvilliorum) suffered heavy losses in Peru and Chile91,115,117, double-crested cormorants (Nannopterum auritum) in North America19,82,108, and great cormorants (Phalacrocorax carbo) in the Baltic region, though with probable limited population-level impacts131. In cranes, hooded cranes (Grus monacha) lost about 10% of their population in Japan and Korea70 while common cranes (Grus grus) suffered high losses in Israel134. Black necked grebes (Podiceps nigricollis) also experienced significant mortality in China137,138 and Canada82.

Mortalities in other wild bird groups

Other wild bird groups, including Anatidae, raptors, and land birds, were also affected with many species reported infected for the first time. Mortality per outbreak was generally lower than in colonial breeding birds due to less gregarious behaviors, but impacts could still have been substantial for long-lived species like raptors.

Anatidae showed the most reported cases (Fig. 2A), with mute swans (Cygnus olor) and geese, especially barnacle and greylag geese among the most reported species in Europe in 2020–2023 with more than 850, 950 and 650 reported cases respectively63,64,65,130,139,140,141,142,143. In the UK, a massive die-off of barnacle geese caused a 32% population loss, while a high percentage of survivors developed antibodies144. In North America, snow geese were particularly affected with more than 18,000 cases reported in 2021–2023, and so were Ross’s (Anser rossii) and Canada (Branta canadensis) geese with around 950 and hundreds of reported cases, respectively19,74,82,109,145. Among dabbling ducks, Eurasian wigeons and mallards were frequently reported in Europe32,40,49,142, while lesser scaups were notably affected in the US with around 1500 reported cases19. In North America, common eiders (Somateria mollissima), a colonial sea duck considered near threatened, experienced severe impacts on the Atlantic coast in summer 2022, with mortality rates reaching up to 18% in a single colony82,108,146.

Raptors were also significantly affected, with common buzzards (Buteo buteo) and peregrine falcons (Falco peregrinus) among the most reported raptor species in Europe, with more than 300 and 70 reported cases, respectively63,64,65,130. White-tailed sea eagles (Haliaeetus albicilla) also experienced notable losses in Norway and Germany147,148, while bald eagles (Haliaeetus leucocephalus) were affected in North America with more than 400 cases19,74,149. Colonial raptors were also affected, including griffon vultures (Gyps fulvus) in France and Spain150 (Fig. 2E), turkey vultures (Cathartes aura) in Chile85,117, and black vultures (Coragyps atratus) in North America with more than 1900 cases19,74. In griffon vultures, immobility was observed, with high juvenile mortality, while many adults survived and developed antibodies150. Great horned owls (Bubo virginianus) were also affected and some showed signs of recovery74,151.

Among land birds, corvids were particularly impacted. Hundreds of large-billed crows (Corvus macrorhynchos) died in Japan152. Similarly, American crows (Corvus brachyrhynchos) were highly affected74, with over 500 cases reported in Canada109. Mortality was also observed in game birds, including common pheasants (Phasianus colchicus) in Finland153 and wild turkeys (Meleagris gallopavo) in the US154.

Ecological drivers of HPAI H5 infections in wild birds

Identifying the mechanisms by which wild birds become infected with HPAI H5 viruses is often challenging. However, certain ecological behaviors, such as migrations, breeding and feeding, can help explain patterns of exposure and transmission. Migratory birds, especially Anatidae, have played a key role in the long-distance spread of HPAI H5 viruses34,60,155. Critical nodes along their migratory routes—breeding, wintering, and stopover grounds—likely facilitated virus transmission through increased interspecies contact, environmental contamination, viral circulation, and reassortments30,53,61,62,66,70,80,132,137,156. However, the significant impacts during summer months, when birds are largely sedentary, and occurrences in non-migratory species suggest that additional regional transmission mechanisms were also at play157.

Certain species, such as gulls and skuas, may have acted as bridge hosts. Their opportunistic feeding or kleptoparasitic behaviors likely increased both their risk of exposure and their potential to disseminate the virus between colonies and into previously unaffected areas99,100,106,107,120,123,158,159. Connectivity between colonies likely further promoted viral dissemination, facilitating inter-colonial transmission in species such as Sandwich terns120,122, great cormorants131, great skuas113, and gannets, with both breeding and non-breeding individuals observed visiting other colonies during outbreaks for this species106,107,110,111,112,158.

Once introduced into dense breeding colonies, HPAI H5 viruses may have spread rapidly due to close contacts between individuals121,123,125,130,132,144. For example, ground-nesting cormorant colonies were significantly more affected than tree-nesting ones, likely due to their higher colony densities131. Decomposing carcasses further contributed to environmental contamination and may have acted as secondary sources of infection121,144,160, as observed in terns and sanderlings6,94.

Predation or scavenging of infected preys appeared to be a major infection route for raptors, given their dietary habits, and is supported by field observations, spatiotemporal clustering of cases and genetic analyses47,70,83,91,131,144,145,147,148,149,151,152,161,162,163,164. This route is also probable in predatory or scavenging seabirds, such as skuas, gulls, frigatebirds, albatrosses, and petrels6,99,100,113,123,125,159, as well as for land birds, notably corvids163,164,165.

Discussion

We identified the major global transmission routes of HPAI H5 viruses via wild birds, which are critical for understanding viral ecology and anticipating where and when viruses might spread, thereby informing future risk assessments. These routes have shown strong alignment with migratory patterns34,38,60, though predicting them remains challenging due to variability across species, seasons and regions. However, certain areas, including North Asia and Europe, have consistently emerged as key viral hubs for both maintenance and onward spread17,18,71,166,167. Proximity to major migratory flyways and critical habitats such as breeding, wintering and stopover grounds has been associated with increased risk of exposure53,155,168. These regions therefore represent valuable targets for enhanced surveillance and may serve as sentinel sites to monitor viral evolution, provided that species-specific migration routes and timings are taken into account169.

Identifying the species responsible for both long-range and local spread remains complex. Dabbling ducks and other Anatidae species have frequently been identified as primary drivers of intercontinental spread13,14,15,16,167. Local maintenance and shorter-distance transmission may have involved other groups of wild birds, such as gulls, but also poultry, notably in high density farming areas170,171 or live bird markets172.

Importantly, this panzootic has affected an unprecedented diversity of wild bird species, including many not previously reported infected by HPAI H5 viruses27, raising serious conservation concerns. Colonial breeding species, such as seabirds or terns, have suffered mass mortality events, resulting in sharp population declines and probable disruptions of colony social structures. Long-lived species with low reproduction rates, restricted geographic ranges or threatened status are particularly at risk, as outbreaks may have caused long-term effects on their populations. The breadth of affected species also raises broader ecological concerns, suggesting an expansion of the virus’s host range and the emergence of novel wild bird reservoirs capable of sustaining viral circulation. This may help explain the unusual persistence of HPAI H5 viruses throughout the year in wild birds in multiple regions although the underlying mechanisms remain poorly understood157. This is a major difference from patterns observed in previous epizootics where different waves were clearly visible. Endemicity was already observed in some countries, notably in South-East Asia, but mainly due to the poultry sector173. A potential future scenario would be the endemicity of HPAI in wild bird populations as a new norm, at a global scale, with sustained circulation157.

Increased viral circulation has also led to repeated spillover events into poultry1,28, and into both domestic and wild mammals3,4,5,6. Both terrestrial and marine mammals have been affected, including large-scale mortality events among marine mammals in South America6,23,174. Reports of probable mammal-to-mammal transmission raise additional concerns about the virus’s adaptation to new hosts, including humans3,4,6,23. A global sustained transmission within mammal populations would be particularly concerning, for both animal and human health and requires a close monitoring of viral mutations. In 2024, repeated spillovers to dairy cattle has underscored even more the propensity of this virus to cross the barrier of species and its zoonotic potential29.

These developments underscore the urgent need for continued research and international collaboration on early detection and rapid response, to mitigate the impacts of HPAI H5 viruses on wildlife, poultry and public health, while enhancing preparedness for future epizootics.

Despite the insights gained through this systematic literature review, many aspects of the epidemiology of HPAI H5 viruses and their relationship with wild bird ecology remain poorly understood. Key questions persist, such as the specific roles different species play in maintaining and disseminating the virus, why certain populations have experienced massive die-offs while others have been relatively spared or why some regions, like Oceania, have so far remained unaffected175. A major limitation lies in the nature of the available data. Current knowledge is largely derived from surveillance networks, which are highly variable in geographic coverage, temporal resolution, and methodological consistency. These networks are further shaped by factors including species detectability, habitat accessibility, conservation status, perceived epidemiological relevance, and levels of public awareness. As a result, substantial sampling biases exist both within and between countries, complicating efforts to interpret findings as representative of broader patterns.

Moreover, only a fraction of wild bird cases is reported in public databases, like WAHIS, and reporting practices vary across regions due to inconsistent case definitions which can be defined at either the individual or event-based level. These discrepancies hinder cross-regional comparisons and contribute to an incomplete understanding of outbreak extent and severity. Standardizing case definition, reporting protocols, and data sharing practices would greatly improve data comparability, transparency, and overall surveillance effectiveness.

Genetic data have proven essential for monitoring viral evolution and reconstructing transmission dynamics through phylogenetic and phylodynamic approaches. However, sequence availability from wild birds remains limited and uneven, due to greater constraints in surveillance and sampling compared to poultry, introducing significant biases and underrepresenting certain host species and regions (Fig. 1). The already complex picture of viral spread176,177, is further complicated by frequent reassortment events and the co-circulation of multiple genotypes25,74,178, making it challenging to clearly delineate transmission pathways.

Mortality and survival data are also frequently incomplete or entirely lacking, especially for mass mortality events, resulting in large underestimation of true impacts179. Even when such data are available, they rarely provide sufficient detail to assess long-term demographic or species-level consequences. More comprehensive studies are needed to evaluate the ecological effects of outbreaks, particularly on vulnerable species to better target interventions180,181,182. Moreover, critical information is frequently reported in grey literature rather than peer-reviewed sources183, meaning that literature reviews alone may fail to fully capture the scale of impacts on wild bird populations. Complementary inputs, such as citizen science initiatives, can also provide valuable insights for both surveillance and the study of ecological impacts184,185.

Predictive modeling could be valuable for exploring future ecological scenarios and assessing potential impacts on specific species, although viral evolution and the development of immunity may hinder the accuracy of such predictions. Our understanding of immune responses in wild birds, and how these shape susceptibility and survival, remains limited, particularly given the wide range of species affected by HPAI186. In this perspective, experimental infection studies can provide critical insights into host susceptibility and immunity, although practical and ethical constraints often limit their use165,187,188,189. Longitudinal studies of seroprevalence evolution in wild bird populations could also contribute substantially to filling this knowledge gap144,190.

Identifying the key wild bird species involved in viral maintenance and dissemination remains particularly challenging. In most studies, infection status and movement data are not integrated101,107,110,112,125, limiting our ability to determine whether —and to what extent— specific species contribute to transmission61,80,81. Continued targeted sampling of live birds is thus essential to detect asymptomatic carriers that may contribute to viral dissemination169. Integrative approaches that combine genetic, epidemiological, ecological and movement data offer promising avenues to gain deeper insights into species-specific roles and transmission routes30,47,71,167. The role of poultry in global dissemination appears limited16 but is much greater at a local scale. Regional transmission and maintenance have been facilitated by high farm density, local connections and imperfect biosecurity measures170,171, as well as through systems such as live bird markets172.

Overcoming these challenges will require strengthened surveillance systems, integrated networks and data infrastructures, as well as enhanced coordination across disciplines, institutions and regions. Such efforts are essential for the long-term prevention and mitigation of HPAI outbreaks in both wild birds and poultry worldwide. Key objectives, underlined by international collaborative expert consultation, include reducing the disease burden in poultry through vaccination and enhanced biosecurity, protecting wildlife and ecosystems by integrating conservation perspectives into monitoring, and safeguarding human health through continuous surveillance of zoonotic spillovers191,192,193. Ultimately, containing the current HPAI panzootic demands a global, multisectoral response within a One Health framework.

Methods

Spatio-temporal analysis of wild bird cases

To assess the impacts of HPAI H5 viruses on wild birds in 2020–2023, we conducted a spatiotemporal analysis of reported cases using data from the WAHIS platform135. Within the study period, three seasons were defined: 2020–2021, 2021–2022 and 2022–2023, ranging from October 1st to September 30th of the following year (extended to the end of 2023 for the 2022–2023 wave), consistent with the seasonality adopted by EFSA128. Due to the limited clade-specific information available in the WAHIS database and the predominance of H5N8 and H5N1 subtypes in reported cases since 2020 (Supplementary Fig. 1)26, we restricted our analysis on these two subtypes and added H5Nx in case of incomplete subtyping. To reduce biases due to reporting inconsistencies across space and time, reports form WAHIS were pooled by months, species and locations. This definition does not account for the quantitative aspects of mortality, notably in mass mortality events. Wild birds were categorized into six groups based on taxonomic, epidemiological, and ecological criteria: Seabirds (excluding Laridae), Laridae, Other waterbirds, Anatidae, Raptors, and Other land birds (Supplementary Table. 1).

Systematic literature review

In parallel, we conducted a systematic literature review following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines (Supplementary Table 3)194. The search was conducted in two databases (PubMed and Web of Science) and was last updated on November 2024. The Boolean query combined two groups of keywords (“wild birds” and “HPAI”), linked by the Boolean operator “AND,” with keywords within each group connected by the operator “OR” (Supplementary Fig. 2). Searches were restricted to titles and abstracts in all databases. To focus on recent HPAI epizootics, the search included only publications from 2020 onward. Only English written articles published in indexed scientific journals were considered. Articles underwent a two-step screening process for inclusion in the final analysis. The first screening assessed titles and abstracts, while the second evaluated full texts. In both screenings, articles had to meet the following criteria: (1) investigate HPAI H5 viruses, specifically the H5N1 or H5N8 subtypes; (2) focus on events occurring from October 2020 to the end of 2023; (3) analyze the spatial spread of HPAI H5 viruses, assess its impact on wild birds or evaluate infection and transmission pathways. Literature reviews were excluded. Primary and secondary screenings were conducted independently by two authors, using a conservative approach in cases of disagreement. To identify additional relevant articles, we employed a snowball sampling technique to include cited articles not captured in the initial search (Supplementary Fig. 2). Data extraction was performed on the remaining articles (Supplementary Table 2), focusing on two key aspects: (1) the spatial spread of HPAI and its links to wild bird movements, (2) infection and transmission pathways in wild birds, as well as the resulting mortality or impacts on their populations.