More than two years after H5N1 bird flu was first detected in U.S. dairy cattle, the virus has proven remarkably good at infecting cows and flooding raw milk with viral material. What it has not done, according to the best available science through early 2025, is learn to spread efficiently between people. That gap is the most consequential question in influenza surveillance right now, and as of spring 2026, it remains unresolved.
Federal investigators, molecular virologists, and structural biologists have each contributed pieces of the picture. Together, their work shows a virus that crossed one species barrier with relative ease and is accumulating mutations associated with mammalian adaptation, yet still lacks the receptor-binding profile needed to move readily through human airways. For dairy farmers, farmworkers, and the agencies tasked with protecting them, the practical stakes are immediate. For the broader public, the risk remains low, but the margin depends on continued surveillance and a virus that has already demonstrated it can surprise.
What surveillance and lab work have confirmed
The foundational facts come from the first federal investigation of the dairy outbreak. A June 2024 CDC report documented that Highly Pathogenic Avian Influenza A(H5N1) had infected U.S. dairy herds, with sick cows showing sharp drops in milk production and other clinical signs. Critically, the investigation found high concentrations of viral RNA in raw milk from affected animals, establishing the mammary gland as a major replication site and raw milk as a significant source of environmental contamination on farms.
That same report described two mild human H5N1 infections in the United States linked to direct exposure to infected or presumed infected animals. Neither case showed signs of onward person-to-person transmission. By February 2025, the CDC’s most recent public risk assessment still classified the immediate threat to the general population as low, noting that confirmed human infections remained tied to close contact with sick cows or poultry.
On the food safety side, the U.S. Food and Drug Administration sampled retail dairy products and evaluated pasteurization, concluding that standard pasteurization effectively neutralizes the virus. The agency issued pointed warnings about raw milk as a potential exposure route, a message that remains the clearest consumer-facing guidance from the outbreak.
By late 2024, the USDA had imposed mandatory H5N1 testing for lactating dairy cattle moving between states, and confirmed infections had been reported in herds across multiple states. The scale of the outbreak underscored how well the virus had adapted to cattle biology.
The mutations driving cattle adaptation
Genomic sequencing has revealed the molecular changes behind that adaptation. A peer-reviewed analysis of dairy cattle in Texas, published in Emerging Infectious Diseases, linked the clinical syndrome in cows to a specific H5N1 lineage and identified the PB2-M631L mutation in the viral polymerase complex, a change associated with improved replication in mammalian cells.
A broader genomic characterization of dairy-associated H5N1 viruses cataloged additional polymerase mutations, including PB2 E627K, a well-known mammalian adaptation marker that appeared in a human case connected to cattle exposure. The authors stressed that these changes reflected rapid within-host evolution after infection, not evidence that the virus had gained the ability to transmit between people.
Controlled infection experiments reinforced the picture. Researchers who inoculated dairy cows with H5N1 and published their results in Nature found that the virus targeted the mammary gland prominently, consistent with the high viral loads seen in milk during field outbreaks. The challenge strain required only a modest set of genetic changes relative to circulating outbreak viruses to establish efficient infection, suggesting the barrier to cattle adaptation was lower than many virologists had expected.
Why the virus still struggles in human airways
If H5N1 has adapted so readily to cows, why hasn’t it made the next jump? The answer, so far, lies in receptor biology. Influenza viruses enter cells by latching onto sialic acid receptors on the cell surface. Human upper airways are rich in α2,6-linked sialic acid, while bird and cow tissues favor α2,3-linked versions. A virus that binds efficiently to α2,6 receptors has a much easier path to human-to-human spread.
Structural work published in Cell in early 2025 showed that the hemagglutinin protein of cattle-infecting H5N1 retains a strong preference for avian-type α2,3 receptors and binds only weakly to human-type α2,6 receptors. Tissue-binding experiments in the same study were consistent with tropism for cow lungs and udder tissue, not human upper airways.
Separate experiments using human bronchus and lung tissue, described in an early release in Emerging Infectious Diseases, found that some cattle-origin H5N1 isolates showed partial affinity for α2,6 receptors, though results varied depending on the assay used. That variability is a reminder that the virus sits in a gray zone: not fully locked into avian-style binding, but nowhere near the human-receptor preference seen in pandemic influenza strains.
The most closely watched experiment came from a Science study highlighted by the National Institutes of Health. Researchers found that the hemagglutinin from a bovine-associated human isolate still preferred avian receptors, but a single amino acid substitution, Q226L, shifted binding toward human-type receptors in the lab. A related study in Nature Communications found a different mutation that broadened binding to α2,3-bearing glycans, essentially deepening the virus’s grip on avian-style receptors rather than redirecting it toward human ones.
The takeaway from these studies is sobering but specific: the cattle-adapted virus is accumulating changes that improve its fitness in mammals, yet the particular receptor switch needed for efficient human airway infection has not occurred in nature. In the lab, that switch can be as small as one amino acid change, which is why surveillance remains urgent.
What scientists and officials are still watching
Several critical unknowns keep this story from being reassuring. The most recent publicly available CDC risk assessment dates to February 2025, leaving a gap of more than a year before this article’s publication. Whether additional mutations have emerged in cattle or human cases during that interval is not reflected in the public record, and the pace of genomic data sharing has been a persistent concern among researchers.
Within-host evolution is another open question. The genomic characterization of dairy-associated viruses documented rapid mutation accumulation after infection, including the PB2 E627K change in a single human patient. How often such changes arise across the broader cattle population, whether they persist when the virus transmits between animals, and what combinations would be needed to cross the threshold into efficient human spread are all unanswered.
There is also the reassortment scenario that keeps influenza preparedness planners up at night. If a farmworker or another person were simultaneously infected with H5N1 and a circulating seasonal flu strain, the two viruses could swap gene segments inside the same cell, potentially producing a hybrid with H5N1’s virulence and a seasonal strain’s ability to spread through human populations. No such event has been documented in connection with the dairy outbreak, but the risk grows with every undetected human infection.
Vaccine development for cattle is in early stages, and no licensed H5N1 vaccine for dairy cows is currently available in the United States. Human H5N1 vaccine candidates exist and some doses have been stockpiled by the federal government, but mass production and distribution would take time if the virus’s behavior changed suddenly.
What this means for farmers, workers, and consumers
For dairy producers and farmworkers, the evidence points to raw milk and direct contact with sick animals as the primary exposure risks. The CDC and USDA have recommended personal protective equipment for workers handling infected cattle, and the mandatory interstate testing program is designed to limit the virus’s geographic spread through animal movement.
For consumers, the message from the FDA has been consistent: pasteurized milk and dairy products are considered safe. The agency’s retail sampling found no viable virus in pasteurized products. Raw milk, by contrast, can carry high viral loads from infected cows, and public health officials have urged consumers to avoid it during the outbreak.
For the public health system, the central challenge is maintaining the surveillance and genomic sequencing infrastructure needed to catch a shift toward human adaptation early. The virus has already demonstrated that it can jump from wild birds to poultry to cattle, accumulating mammalian-adaptive mutations along the way. Each new host species gives it another evolutionary laboratory. The fact that it has not yet acquired efficient human transmissibility is not a guarantee; it is a window, and the width of that window depends on how closely scientists and regulators are watching.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
