I think it’s fair to say H5N1 becoming seemingly endemic in US dairy cattle in the past year caught us off guard. The virus has spread widely within and between US dairy herds, has caused mild infections in a number of people in close contact with infected cows, has killed a lot of cats on farms (and a few from drinking raw milk from infected cows)… and it isn’t likely to go away any time soon. 

Dairy veterinarians are one of the higher risk groups for exposure from infected cattle because of their close and frequent contact with these animals, particularly when cattle are ill. Surveillance testing of people at high risk for exposure to H5N1 influenza can help us get a handle on how much (if any) under-the-radar cow-to-human transmission may be happening, so it was great to see the release of the results of just such a surveillance study in dairy veterinarians in the US (Leonard et al. MMWR 2025).

In this study, researchers tested blood samples from 150 veterinarians with cattle contact and tested them for antibodies against H5N1 influenza. The presence of antibodies would indicate previous infection, whether or not the person was ever sick from the virus.

Three of 150 (2%) dairy veterinarians were positive for H5N1 flu antibodies, but none of those reported having had signs of illness that could have been attributed to flu, and none reported working with dairy cattle that were known to have been infected with H5N1 flu. If that’s accurate, it could indicate a few things, including possibly:

• working on farms where there was mild disease in cattle from H5N1 flu that was not recognized
• working on farms where there was disease in cattle from H5N1 flu but the cattle were not tested for it
• veterinarians were exposed to H5N1 flu in some other way from animals or the environment, such as through contact with other animals or raw milk. (One of the veterinarians who tested positive also had contact with infected poultry, so that’s another potential source of exposure)
• veterinarians were exposed through unrecognized human-to-human spread of H5N1 (which would be the most concerning possibility)

One of the seropositive veterinarians worked with dairy cattle in Georgia and other cattle in South Carolina. Neither of those states is known to have H5N1 in dairy cattle (see map above from the report), but the degree of surveillance and (more importantly) reporting is variable across the US. This would suggest that testing of cattle in Georgia needs to be ramped up to see if they have unrecognized infected herds.

All three antibody-positive veterinarians “reported wearing gloves or a clothing cover when providing veterinary care to cattle (including a variety of clinical activities, such as pregnancy checking or surgery)”. That’s strange wording, since those are two distinctly different types of PPE. Virtually every dairy veterinarian is going to wear coveralls (a clothing cover) on farm, so that stat tells us nothing about how many of them wore gloves (nor whether glove use may have been suboptimal, as it often is on farm). There was no use of eye or respiratory protection, which is far from surprising and something the veterinary profession needs to improve, as we do a poor job of using respiratory protection and rarely use eye protection when we’re dealing with animals with respiratory infections, even when they could be zoonotic.

The fact that all three antibody-positive veterinarians reported no obvious consequences of H5N1 flu infection is good news on many levels. However, asymptomatic infections raise some concerns, since if people are asymptomatically infected but still infectious, it may allow the virus to spread silently through the population, at least for a while. We have no idea if infected people shed the virus at levels that can infect others, but it’s something for which we need to be on the look out.

Any H5N1 flu infections in a person is bad, because of the potential for severe disease in the person and, even more importantly, the potential for evolution of the virus to transmit more easily among people. The more H5N1 encounters humans, the more opportunity it has to become adapted to humans. Infection of people concurrently infected with human flu virus strains creates opportunities for recombination of both viruses, which can lead to rapid and significant undesirable changes and emergence of new strains.

This is far from a doomsday report, but it highlights some things that we need to keep watching. It also shows why we need more effort to contain the spread of H5N1 flu in domestic animals. The data here are a bit limited, but they’re an important step in our understanding of this virus. A parallel study of the general population would complement these data, as would more focused study of veterinarians and farmers from affected farms, and veterinarians working with other species.