Testing air and cages in poultry markets catches deadly bird flu that swabbing birds misses

The standard way to check for dangerous viruses in a live poultry market involves catching a bird, holding it still, and swabbing its throat or cloaca. It is slow, labor-intensive, poses safety risks to workers, and misses infected birds that happen not to be sampled. In markets where hundreds of birds cycle through daily, testing a handful provides limited information.

A research team at Duke-NUS Medical School in Singapore has demonstrated a more effective alternative: skip the birds and sample the environment instead. By collecting air samples, swabbing cages, and testing water used in poultry processing, the team detected a broader range of viruses -- including highly pathogenic avian influenza H5N1 -- than traditional bird-by-bird testing could find.

Forty viruses from air, water, and surfaces

Between January 2022 and April 2023, researchers collected environmental samples from two live poultry markets in Cambodia. Rather than targeting specific pathogens, they used metagenomic sequencing, a technique that identifies all viral genetic material present in a sample without pre-selecting what to look for.

The approach detected genetic material from 40 different poultry viruses, including multiple influenza strains and coronaviruses. Air samples consistently captured the greatest diversity. When the researchers compared environmental results with traditional swabs taken from chickens and ducks at the same markets, the discrepancy was striking: highly pathogenic H5N1 showed up in environmental samples even when concurrent bird tests found nothing.

"We showed that direct animal testing is not always necessary to detect pathogenic viruses in live-bird markets," said Dr. Peter Cronin, first author of the study. "Sampling air, water, cages and surfaces can reveal a wide range of poultry viruses, including avian influenza, even when those same viruses are not detected in the birds at the time."

Why birds alone are not enough

The finding is not as counterintuitive as it might sound. A virus circulating through a flock sheds into the shared environment -- onto cage surfaces, into the air through respiratory droplets, and into water used for washing and processing. That environmental reservoir accumulates viral material from many birds over time, providing a cumulative signal that testing a few individual animals cannot match.

Some of the H5N1 strains detected belonged to genetic lineages known to pose significant risks to both poultry and humans. Their presence in air and surface samples, but absence in concurrent bird swabs, suggests that relying solely on animal testing may substantially underestimate the true level of viral circulation in these markets.

The team also found that air samples collected near slaughter and holding areas contained viral material from multiple poultry pathogens simultaneously. Workers and customers in these areas may be exposed through shared air, a finding that underscores the importance of ventilation and market layout in managing zoonotic risk.

Complement, not replace

The researchers are careful to note that environmental surveillance should work alongside traditional animal testing, not supplant it. Some viruses -- particularly those carried by ducks, which were present in smaller numbers at the markets studied -- were more reliably detected through direct swabbing. The most comprehensive surveillance strategy combines both approaches.

"This study provides a more comprehensive view of viral circulation in live poultry markets than is possible through single-animal testing alone," said Prof. Gavin Smith, co-senior author and director of the Emerging Infectious Diseases Signature Research Programme. "By applying unbiased metagenomic sequencing to environmental samples, we capture viral material shed across shared air and surfaces, enabling broader detection in a cost-effective and scalable manner."

Scaling the approach

Live poultry markets are widespread across Asia, providing fresh food and supporting livelihoods for millions. They are also recognized as high-risk interfaces where animal and human populations mix closely, creating opportunities for viral spillover. Effective surveillance at these sites is a cornerstone of pandemic preparedness.

The Duke-NUS team is now exploring how environmental surveillance can be applied in other settings, including pig slaughterhouses and wildlife environments. The approach is inherently scalable: collecting air samples and surface swabs requires less specialized training than handling live animals, carries lower occupational risk, and can be performed more frequently at lower cost.

The study does not address whether environmental sampling can detect early-stage outbreaks quickly enough to trigger public health responses before human infections occur. That question -- essentially, whether the method works as an early warning system rather than just a more sensitive census -- would require prospective studies with real-time reporting. But as a surveillance tool, the environmental approach fills a clear gap that bird-by-bird testing cannot.