As H5N1 avian influenza continues its unprecedented spread internationally among wild birds, domestic bird, different mammals and even people, concerns continue to increase. One of the things that increases risk with influenza viruses is the number of species a strain can infect. The more species the virus can infect, the more risk of exposure to others, and the more risk of the virus changing (for the worse) within a new host. Infection of different mammals gives the virus more opportunity to adapt to and infect other mammals (and humans are included in that group). Some species can also be “mixing vessels” for influenza viruses. If an individual is infected with two different flu viruses at the same time, there’s risk the viruses can get mixed up together and create a new strain.

A recent paper in Emerging Infectious Diseases about the potential for H5N1 flu to infect horses raises some interesting questions, and has resulted in some predictable over-reaction.

The study (Damdinjav et al. 2024) looked at blood samples from horses from two parts of Mongolia: a wetland with a lot of birds, and a dry area with lower bird density. They studied 24 horse herds by testing 10 animals per herd, three times a year between July 2021 and October 2023. There was no evidence that there had been more or any unusual illness in the horses in these areas during the sampling period. They collected a total of 2160 equine blood samples.

• Nine samples (0.4%) were positive for antibodies against H5, with the assumption that H5 positives meant exposure to H5N1 influenza. That’s a reasonable assumption, but there’s no guarantee as there could be other H5 influenza strains in wild birds in the area as well (however they did rule out cross-reaction with equine flu strains).
• A further 8 samples (0.4%) were “doubtful.”
• The rest of the samples were negative for H5 antibodies (indicating no evidence of an immune response to an H5 influenza virus).

They took the 17 positive or doubtful samples and tested them using an H5N1 viral neutralization assay.

• Only 2 samples were positive using the H5N1-specific test (that’s less than 0.1% overall).

I’m always a bit wary of low-prevalence serological studies. Even if the test was 99% specific (99% of positives are true positives), the low prevalence reported here could be within the error rate. It’s an inherent issue with serological tests where we don’t have (or at least can’t be certain of) 100% specificity. I’m not saying the results are wrong, but we have to consider that some or all of the very small number of positive results could be false positives. More study is needed to know for sure.

If these were true positives, the very low prevalence is important to note. It would strongly suggest that H5N1 is not transmitted between horses.

• If a new flu strain gets into a population, I’d expect really rapid, widespread transmission. A large percentage of horses would be infected if there was horse-to-horse transmission.
• The low prevalence suggests sporadic dead-end spillover infections that were likely all bird-to-horse.

What does this tell us?

• That a small number of horses in an area where H5N1 was circulating were infected with H5N1. Interesting and noteworthy, but that’s it.

What doesn’t this tell us?

• Whether horses were infected with a currently circulating strain of H5N1 influenza.
• Whether horses got sick from H5N1 influenza infection.
• Whether horses can spread H5N1 influenza (to other horses or to other species).

The authors concluded that horses are susceptible to H5N1 infection. That’s probably true, but infection may or may not causes disease. They also suggested that “spillover events are likely frequent,” which seems to be a stretch based on these data.

What’s the risk of spread of H5N1 influenza to people from horses?

Presumably it’s really low. There’s no evidence that horses are a realistic source of H5N1 infection. As discussed above, the low prevalence is strongly suggestive that there was no transmission between horses. We should keep looking, but at this point, nothing suggests to me that horses pose a realistic risk to people or other animals.

What about the risk of recombination of flu viruses in horses?

Horses have their own influenza strains, mainly H3N8 equine influenza. So, concurrent infection with H5N1 could create a recombination risk. But it’s probably much more of a concern for horses than for people. H3N8 equine flu is well disseminated internationally and has been for decades, but horse-to-human transmission of H3N8 is exceptionally rare. So, if H3N8 combines with another virus that has a poor ability to infect people, it’s unlikely that the new virus would be more likely to infect people. We’re more concerned about H5N1 recombining with a human flu virus or animal flu virus that more readily infects people (like some swine H3N2 strains). Recombination of equine H3N8 and avian H5N1 could be a big deal for horses, though, if it becomes a strain that can cause serious disease in horses and for which horses have little immunity (because it’s different enough from the H3N8 their used to).
That said, while the risk may be greater for horses than humans, we don’t want more circulating flu strains; the more infections there are in horses, the greater the risk of that virus changing even more.

Take home message

The risk of H5N1 influenza infection in and from horses is probably very low. Horses rank fairly far down the list of species about which I’m concerned.

However, they’re still on the list. While I don’t have any plans for flu surveillance in horses around here (no available funding being the main barrier), I’m definitely paying attention to them. If we hear about more or atypical disease in horses, especially if it coincides with more flu activity in birds in the area, I’ll want to start testing. The issue with H5N1 in dairy cattle in the US caught us off guard, which shows why we have to remain vigilant and be ready to act quickly and aggressively. (The lack of quick aggressive action in the US with dairy cattle was likely a major reason this virus has spread so widely there.)