We collated and jointly analyzed information from griffon vultures collected from Spain and France in the framework of several research and surveillance programs dedicated to the study of vulture ecology, population dynamics, or HPAI outbreak follow-up (Appendix). In addition, the sampling included birds from national and local surveillance programs or those submitted for diagnosis to the institutions of the authors. In Spain, griffon vulture captures were performed in numerous colonies across the Iberian Peninsula during 2020–2022. Adult griffon vultures were captured with net traps or walk-in traps at vulture feeding stations, and nestlings were captured at nests. In France, 3 capture sessions were performed in 2022 and 2023 using walk-in traps at 5 sites.

All birds were ringed and a subset of vultures were fitted with GPS satellite transmitters. Blood samples, as well as oropharyngeal and cloacal swab samples, were collected from live birds. From dead birds, vascular feathers and tissues from the main organs (spleen, pancreas, heart, brain, trachea, intestine, lungs, and liver) were collected.

We extracted nucleic acids and submitted them to generic matrix gene qRT-PCR. We then submitted positive samples for nanopore sequencing (Oxford Nanopore Technologies, https://nanoporetech.com) and used consensus genomes in molecular marker and phylogeographic analysis.

We separated serum from the cell pellet by centrifugation in blood samples and stored at −20°C until analysis. We inactivated and tested serum samples with commercial ELISA and submitted positive samples to a hemagglutination inhibition (HI) test. We categorized HI samples as positive for an antigen if the titer was >16. We then estimated true seroprevalence by using the epidemiologic calculator Epitools (https://epitools.ausvet.com.au/prevalence).

We examined movement patterns of GPS-tagged griffon vultures during the 4-month period from March 1–June 29 during 2022 (outbreak year) and 2023 (control year), characterizing tracks as local or transit movements or immobility. We calculated daily distances traveled (DDT) and used generalized linear mixed models to identify factors affecting DDT. We first examined the movement patterns of a large sample of 114 birds, then examined the spatial and temporal dimensions of the movements of a subsample of 16 birds that remained in Europe during the study period.

HPAI H5N1 infections were first detected in griffon vultures in southern Europe in spring 2022 in Spain and France. Infection led to central nervous clinical signs as well as reduced activity, immobility at the roost or nest, and death in some adults. Deaths were also recorded in nestlings, either from direct effects of infection or lack of parental care (19).

We investigated the origin of infection and assessed the nature and magnitude of viral spread through an integrated analysis of virologic, serologic, genomic, and ecologic data obtained from field sampling in France and Spain. In addition, we evaluated potential sources of exposure, studied population connectivity pathways, and investigated the potential of long-range contamination by movements of infected birds using long-term GPS tracking data.

Serologic results reflected the circulation of H5 HPAI among flying birds of all sampled colonies of griffon vultures. The high ELISA seroprevalence observed in flying birds seems to indicate widespread transmission but high survival within the meta-population, as opposed to the dramatic mortality rates observed in Sandwich terns, northern gannets, and bald eagles (20,23,24). In contrast, ELISA seroprevalence was considerably lower in nestlings; seropositivity was detected in just 1 colony (Appendix Table 1). That lower seroprevalence could be because of a reduced exposure of nestlings or the result of a high mortality rate. Those seropositive nestlings, detected at the same time as H5N1 AIV–mediated deaths of nestlings in neighboring nests, fledged successfully, demonstrating that infected nestlings, as well as infected adults, were able to survive the infection.

The spread of the virus through the griffon vulture populations in Spain and France was very fast; most H5 HPAI PCR–positive dead birds were collected during April–June 2022, and almost all live and dead birds sampled in both countries after that period tested negative for H5 HPAI virus by qRT-PCR. None of the antibody-positive vultures from our sample set tested positive by PCR, suggesting the absence of active HPAI H5N1 circulation after June 2022, which might relate to the clearance of infection by surviving vultures (25,26). Sandwich terns sampled in 2022 and 2023 showed a similar pattern; seropositivity was detected in adults in the absence of viral shedding (24). However, the lack of data regarding persistence of AIV immunity in griffon vultures, which is known to be highly dependent on the species and age of the host, as well as the subtype of virus, does not preclude circulation at a low prevalence or in the absence of clinical signs (27,28). Presence of H5-specific AIV antibodies as much as a year after the outbreak, albeit at low titers (Appendix Table 2), might be from antibody persistence or reexposure or from exposure to a different H5 AIV. In fact, AIV seropositivity of vultures in Spain to AIV other than H5 before 2022 provides circumstantial evidence that exposure of vultures to AIV could occur occasionally, which contrasts with the common assumption before the H5N1 outbreak that vultures were either not exposed or not susceptible to AIV infection.

Of note, HPAIV-related deaths of bearded vultures (G. barbatus) were reported concurrently to the griffon vulture outbreak (SM17). In contrast, no evidence of infection was found in 2 other cooccurring species of vultures, namely cinereous vultures (Aegypius monachus) and Egyptian vultures (Neophron percnopterus), which also regularly feed alongside griffon vultures (data not shown).

Phylogenetic and phylogeographic analyses of H5 genetic sequences obtained during the HPAI H5N1 outbreak offered key insights into the potential origins and transmission dynamics of viral lineages among griffon vultures. Of note, the inclusion of almost all griffon vultures within the same genetic cluster, despite having been sampled from geographically distant locations in Spain and France, suggests that they could have shared a common source of infection because of their scavenging feeding behavior, or that some could have been initially infected and subsequently transmitted the virus to others.

Those analyses enable us to draw hypotheses regarding the origin of the virus that infected vultures. This virus was genetically close to strains found in greylag geese and other wetland species, such as grey herons and white storks. Griffon vultures rarely forage in wetlands in France, but they do more often in southern Spain, where they can feed on livestock or wild mammal carcasses in marshes (i.e., potentially close to waterbirds) (Appendix Figure 5). Interactions with waterbirds might also have occurred at small waterbodies in southern Spain, where griffon vultures regularly bathe. Contacts could also have occurred in open urban landfills in northern Spain, where griffon vultures regularly feed alongside other wild bird species that are highly susceptible to the virus, such as gulls and white storks (29–32). Previous studies have shown potential transmission of low pathogenic avian influenza viruses between species frequenting the same landfills (33,34). Finally, contacts might involve opportunistic carnivorous mammals, such as red foxes (Vulpes vulpes), in which deaths from HPAI H5N1virus have been reported throughout Europe (35). Uncertainties remain regarding the specific mechanisms of the griffon vultures’ contamination.

Our phylogenetic and phylogeographic analyses further suggest that the introduction of HPAI in griffon vultures from poultry farms seems unlikely. This conclusion is also supported by the behavior of griffon vultures, which do not typically visit poultry farm premises. Moreover, culled poultry from affected farms are discarded under strict biosecurity regulations, making the contact of griffon vultures with infected dead poultry unlikely. However, we cannot rule out that griffon vultures might have accessed inadequately discarded undiagnosed dead backyard poultry in some regions of Spain that could also have been consumed by gulls and storks, leading to the further spread and detection of the strain in both wild waterbirds and vultures (31,32).

The limited genetic diversity of the virus observed in griffon vulture populations, contrasting with the wide geographic distribution of infected birds, suggests that the virus spread in the southwestern Europe metapopulation through intraspecific contamination. The ecology of griffon vultures, especially their feeding behavior and their colonial nature, could explain this finding. Dense short-term aggregations during feeding on carcasses or at vulture feeding stations and dumpsites, where hundreds of individual birds congregate, make griffon vultures particularly vulnerable to airborne pathogen transmission (36,37). Subsequently, movements of infected birds over long distances could easily have contributed to virus dissemination to the whole population (38).

Phylogeographic reconstructions reveal a spatial dissemination pattern across 4 distinct regions, originating from central Spain, spreading to France in the Massif central and the Alps, and subsequently disseminating to the western part of the Pyrenees and southern Spain. This finding is coherent with the analysis of telemetric data, which show an overlap in the distributions of several GPS-tagged birds in Spain and France and long-range movements occurring between populations, particularly in spring, concurrent with the 2022 outbreak (Appendix). Such movements were also observed in other years (39), and it thus seems unlikely for them to have been triggered by the outbreak, as observed with northern gannets (23,40,41).

The nestling that was found seronegative and without clinical signs but tested positive for HPAI H5 in the vascular feather provides circumstantial evidence for shedding during the presymptomatic period, because shedding from feather follicles has recently been described as an efficient route of HPAI transmission (42). Experimental infection of red-legged partridges (Alectoris rufa) with an HPAI H7N1 virus evidenced an incubation period of 3 days with shedding from day 1 (43), whereas a similar approach evidenced a 5-day incubation period in falcons (Falco spp.) experimentally infected with HPAI H5N1 (44). Thus, under the hypothesis of viral incubation lasting >3 days, a griffon vulture infected in the Pyrenees would have enough time to reach southern Spain or the French Alps before showing clinical signs and reduced mobility (19). As an example, an immature vulture (Imm_FR_JOR) traveled from southern Portugal to the French Alps in 6 full days (Appendix).

The outbreak described in this study appeared to have had only a mild effect in terms of the conservation of griffon vultures. Compared with long-lived seabird populations in which a large proportion of adults died, mortality in griffon vultures mostly affected nestlings and only few adult birds; adult survival is the most sensitive demographic parameter in such a long-lived species (45). In addition, the outbreak struck the world’s largest population of griffon vultures, which could withstand such an ephemeral reduction in breeding success. However, even if griffon vulture populations seem able to overcome this HPAI outbreak, they face multiple threats on a global scale, particularly poisoning and persecution (46–48). The introduction and circulation of a new infectious pathogen could add additional pressure on the population. Unfortunately, the consequences of a population collapse of necrophagous birds could be catastrophic, especially from a sanitary point of view, in that longer persistence of the carcasses they eliminate would increase risk for pathogen persistence and spread in the environment (17,49).

Despite a likely limited epidemiologic role of griffon vultures in the circulation of HPAI in Spain and France, with very low permeability between griffon vulture populations and poultry farms, the infection of this new compartment raises multiple questions. In particular, this outbreak demonstrates the ability of this virus (and potentially other highly contagious pathogens) to spread rapidly through a population after a single introduction and shows that even a rare event has the potential for devastating effects.

In conclusion, the recent evolution of HPAI H5N1 has led to this pathogen being considered a severe concern for endangered bird species, especially those with colonial and scavenging behavior. Integrating the epidemiology of the virus with the ecology of the host species is key to a better understanding of outbreak dynamics and possible effects on wildlife conservation. More generally, implementing a multidisciplinary approach will be necessary to overcome these new challenges.