Study finds strong evidence that bird flu spreads through the air between farms

By Tarun Sai LomteReviewed by Susha Cheriyedath, M.Sc.Feb 24 2025

New research uncovers how the wind may have carried H5N1 between farms, challenging previous assumptions and reshaping outbreak prevention strategies.

Study: Genetic data and meteorological conditions: unravelling the windborne transmission of H5N1 high-pathogenicity avian influenza between commercial poultry outbreaks. Image Credit: Handy Yu / Shutterstock

*Important notice: bioRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information.

In a recent study posted to the bioRxiv preprint* server, researchers provided genetic evidence suggesting the feasibility of windborne transmission of the highly pathogenic avian influenza (HPAI) H5N1 virus. Although windborne transmission of avian influenza has been theorized, documented cases remain rare, and previous air-sampling studies have generally detected the virus only up to ~150 meters from infected farms. This study provides strong evidence supporting long-distance transmission, though it does not claim to definitively prove it.

Poultry farms represent significant air pollutant emitters. The pollutants vary in size, shape, and origin but generally comprise liquid droplets, gases, and inorganic and organic matter. During an HPAI outbreak, viral particles could be detected in the air and dust inside poultry markets or farms. While airborne transmission is the primary infection mode in poultry, contaminated air could contribute to disease transmission. The study also underscores that mechanically ventilated poultry houses may increase exposure risk by drawing in contaminated air from the surroundings, potentially facilitating infection even at low viral concentrations.

Wind has long been implicated in long-distance transmission of influenza A virus (IAV), albeit field reports on this transmission mode are scarce. Nevertheless, sampling studies have detected IAV up to 150 meters away from infected farms, with low viral loads and positivity rates, suggesting that viral load is inversely related to distance. However, under specific meteorological conditions, windborne transmission over longer distances may be possible.

The study and findings

In the present study, researchers showed that windborne transmission of the HPAI H5N1 virus is possible over a distance of 8 km. The study carefully reconstructs the sequence of outbreak events using genetic, epizootiological, meteorological, and landscape data, aligning viral spread with wind direction and speed. On February 4, 2024, a sudden increase in deaths occurred in a large farm (B) of 50,000 fattening ducks in the Czech Republic, affecting around 800 ducks. Two days later, deaths surged to 5,000 ducks. The farm was located near a lake accessed by wild birds.

The flock was depopulated to control the outbreak, and 10-km surveillance and 3-km protection zones were established. However, on February 12, 2024, an HPAI outbreak occurred in two chicken farms (C1 and C2) within the surveillance zone. C1 and C2 comprised two houses each, with birds housed in indoor cages without litter.

Given their location inside the surveillance zone, a nutritional supplement was administered on February 7 to boost immunity. Notably, in the week before HPAI onset, both farms gradually reduced their feed and water consumption. This decline, initially attributed to the supplement, aligns with the slow disease progression and prolonged incubation period described in the study. HPAI was first detected in houses A and B of C1 (C1A) and C2 (C2B), respectively, with a slight increase in deaths.

Notably, infection and deaths in affected houses occurred in areas proximal to air inlets. This spatial pattern suggests that contaminated air entering through ventilation systems played a role in disease spread. C1A was depopulated to protect the indigenous gene pool in C1B, which housed an important genetic reserve of around 2,000 birds. However, C1B ultimately required depopulation. Conversely, C2A was unaffected throughout.

The researchers leveraged whole-genome sequencing (WGS) and phylogenetic analysis to examine the relationships between affected farms. Overall, they acquired 38 H5N1 genomes; nine were obtained from B, five from C2B, three from C1A, 15 from C1B, and six from backyard poultry near B. All H5N1 strains were of the DI genotype, forming a common subclade. Three strains from B showed 100% nucleotide identity to C1A strains and the index C1B strain.

By contrast, later strains from C1B and C2B exhibited farm-specific clustering. Further, strains from the backyard poultry near B had a close relationship with C1A, index C1B, and B strains. Median-joining network analysis showed that all but one strain from B and early C1 strains formed a central node, from which three branches emerged. Two branches were clusters specific to C1B and C2B, while the third branch represented strains from the backyard poultry.

This genetic overlap suggested a common origin, with B being the likely source of spreading the disease to C1 and C2. Further, duck and chicken farms were located 8 km apart and were active in distinct markets. Interviews with the general manager and veterinary inspector indicated no interactions between B and C.

Infected Premises. Layout of affected houses on farms B (A), C1 (B), and C2 (C). On C1 and C2, tunnel ventilation system airflow is indicated with orange arrows for inflow and blue arrows for outflow.

As such, the possibility of transmission associated with humans from B to C1/C2 or C1 to C2 was excluded. Additionally, the study ruled out other possible transmission routes, including wild birds, rodents, and insect vectors. The enclosed design of the chicken farms and the absence of significant water bodies in the region further reduce the likelihood of wild bird-mediated transmission.

All farms were located at comparable altitudes: 408–448 m. Prevailing weather conditions were minimal precipitation, intermittent sunshine, and extensive cloud cover, with humidity averaging 77%–81%. Temperatures were not below 6 °C, except on the night of February 7 and 8. Wind data revealed a key transmission window: from noon on February 4 to midnight on February 5, when continuous winds from the southwest or west, with speeds reaching up to 10 m/s, aligned with the inferred infection route.

Conclusions

The HPAI H5N1 outbreak began on February 4, 2024, at the duck farm. While the source of this outbreak was unknown, mallards at a nearby lake were the most likely origin. That a proximal backyard flock was infected with genetically similar viral strains suggested that wild birds sustained viral circulation in the region.

Aligning the timeline of events with wind speeds and direction supports a transmission window of February 4 to 9, with peak transmission likely occurring from February 4 to 5. This was supported by 1) the slow progression of disease with a protracted incubation period characterized by reduced feed and water intake before symptom onset, 2) the location of affected birds in areas closest to air inlets, and 3) the genetic identity of viral strains in donor and recipient farms.

In addition, field investigations ruled out other possible infection routes during this period. Importantly, the study challenges the long-held belief that dust generated during depopulation is the main airborne transmission risk. Instead, it suggests that finer aerosolized viral particles, carried by the wind while the infected flock was still alive, may be more significant carriers.

Therefore, index C1B and C1A viruses were windborne seed strains from B, which evolved locally and formed a farm-specific cluster. Overall, the combination of meteorological, landscape, epizootiological, and clinical factors suggests that windborne transmission was the most plausible explanation for viral spread in this case.

These findings highlight the need to consider windborne spread in future outbreak mitigation strategies, particularly in farms with high-density poultry populations and mechanically ventilated systems that may enhance exposure.

*Important notice: bioRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information.