Let’s start with the totally unsurprising.

One cat was from an infected dairy farm.

• Duh. That one’s easy to explain.

Three were “known indoor/outdoor cats that hunted mice and/or small birds as prey and also spent time indoors with their owners.”

• These aren’t too surprising either. When the virus is circulating in birds, species that prey on birds are likely to get exposed.
• Genomic evaluation of the virus will help tell us if the strains in these cats are consistent with those circulating in local birds (as they should be), versus the slightly different strains circulating in cattle (hopefully not).

Now to the unexpected cases. These are the most interesting and important cases, and they illustrate some major gaps in our understanding and approach to this virus.

Two cats were indoor cats with no known exposure to H5N1.

• These cats didn’t have contact with birds or dairy cattle.
• Presumably they also didn’t have contact with any person known to have had H5N1 (as there are still few enough of those cases) or contact with high risk people like dairy or poultry farm workers, but that’s not explicitly stated.
• I also presume the authorities took a good enough history to make sure “indoor” really meant “indoor,” since “indoor” is sometimes an aspiration for cats, not a lifestyle. (Veterinarians are all too familiar with “indoor” cats that have been hit by a car or gotten into a fight with some kind of wildlife.)

So how did two “indoor” cats get exposed to the H5N1 virus?

A few potential, but still unlikely, possibilities come to mind:

• Close contact with aerosols from infected birds or bird feces through window screens?
• Close contact with aerosols from infected mammals (e.g. other cats) through window screens?
• Birds that snuck into the house?
• Unidentified infected humans?
• Contaminated material from outside tracked in on someone’s shoes or clothes?
• Raw food? (Including possibly raw milk?)

The latter, in the form of raw diets, has been shown to be a risk factor in a few outbreaks in cats, so we can’t dismiss it. However, the odds of a raw diet containing meat from an infected bird in Colorado are really low. So, I think potential causes are still wide open at this point. Genomic analysis should help determine if the virus from these cats is most consistent with the strain in dairy cattle or circulating avian strains.

These cases also highlight something else: surveillance bias. If you don’t test, you don’t find. If testing is focused (or restricted to) cats with known high risk contacts, we can get into a self-fulfilling prophecy of “cats only get infected if they have risk factor X.” The two unexpected cases in indoor cats show that we might need to throw a wider surveillance net, both to find more infected cats and to understand how this virus is being spread.

The disease presentation is also important. Five of the six cats had “an initial complaint of lethargy and inappetence, followed by progressive respiratory signs in some and fairly consistent progressive neurologic signs in most.” It’s not clear how disease progressed in the sixth cat, or if it was perhaps found dead. Consider that if testing focuses on cats with neurological disease, we’ll bias ourselves to thinking that this virus always causes neurological disease. I’m not sure at this point whether H5N1 infection usually causes serious neurological disease or whether the cats with serious neurological disease are just the small subset that we test for H5N1. If those are the only cats we test (perhaps as a secondary test when the cat was a rabies suspect and has tested negative for that virus first), then we’re not going to understand the true picture.

We need better and broader surveillance. I get worried when the focus of testing is on a narrow population. 

-

The risk posed by cats to people and other animals is completely unclear at this time. Some earlier data suggested that cats could have pretty high viral loads in respiratory secretions, so I think we have to assume that infected cats pose some transmission risk. That doesn’t mean we should panic or not try to treat them, but we should make sure we use good infection control practices around suspected and confirmed cases.

Information on human-to-pet mpox transmission is still very limited.  https://wwwnc.cdc.gov/eid/article/30/10/24-0632_article - Two Fortunately, we now have some new information from the US. An ahead-of-print article in Emerging Infectious Diseases ( https://wwwnc.cdc.gov/eid/article/30/10/24-0632_article - Morgan et al. 2024 ), describes an investigation in the US between July 2022 and March 2023. They collected skin swabs and blood samples from pets of people with mpox from 21 households. They sampled 24 dogs, 9 cats and 1 rabbit, obtaining a total of 191 swabs from animals and 56 swabs from the animals’ environments (e.g. beds, bedding, toys, bowls). The results were interesting:

• MPXV DNA was identified by PCR from five animals (4 dogs, 1 cat) from 4 households. Two positive dogs were from the same household. That’s an overall positivity rate of 17% in dogs and 11% in cats sampled.
• Multiple positive results were obtained from all of the positive animals. Overall, 12% (22/191) of animal samples were positive.

Breaking it down by body site, positives were from:

• 29% (4/14) of skin lesion samples
• 16% (6/37) of ventral skin/fur (stomach) samples
• 12% (4/33) of dorsal fur (back) samples
• 11% (4/35) of periocular (around the eyes) samples
• 8% (3/36) of anorectal (butt) samples
• 3% (1/36) of oral samples

Researchers also tested for RNase-P as an indicator of human DNA contamination, and all PCR-positive samples were also positive for RNase-P. While that doesn’t mean that the MPXV DNA all came from direct human contamination, it means there was evidence of human contact with the same sites at the same time.

Seven animals had skin lesions, but five of those had skin lesions before the owner got sick, so those cases were likely unrelated. Two pets developed skin lesions after the owner got sick (or at least no one noticed them before that) and both were PCR-negative for MPXV.

Twenty-nine percent (29%) of households had positive animal-associated environmental samples (e.g. dog beds, toys, bowls). All households with positive animals had positive environmental sites, but so did two households without positive animals. Environmental samples are always hard to assess in situations like this, because a positive result could be from contamination from the person or the pet. Environmental samples also don’t tell us risk, since PCR testing will also detect dead virus. It’s more of an indicator of the viral load in the area, and provides some suggestions of contamination routes and what sites might potentially be of concern.

PCR is a really sensitive testing method that can pick up even small bits of dead virus, so the researchers also followed up with viral culture, which is more time consuming, specialized and hazardous, while also being less sensitive – which is why it’s not typically uses it for initial testing. Culture is useful for figuring out if there truly was an active infection, because it only detects live virus. All PCR-positive samples in this study were negative on culture. That’s a bit hard to interpret with confidence.

• It could mean that infection levels were low, since culture isn’t sensitive.
• If sampling was done late in infection (which would be common in studies like this, since researchers can’t usually get into households really quickly, then active infection may have waned by the time of sampling.
• It could be that samples didn’t pick up much virus.
• It could also mean that PCR-positives were from contamination with dead virus.

Testing for MPXV antibodies in the blood was also performed in 20 dogs at the time of the first site visit, and all were negative. More importantly, 22 dogs were tested 3-4 months later and still had no detectable antibodies. That’s more important because initial negative results might just be because the body hadn’t yet produced detectable antibodies to the virus, but by 3-4 months, antibodies should be present if the animal was previously infected. It’s not a complete guarantee that the dogs weren’t infected, since we don’t know how good the antibody test is or how robust antibody production might be if infection was mild and transient. However, it provides more evidence that PCR-positive results could be due to contamination or a very low (and maybe irrelevant) level of transient infection.

Does this mean dogs and cats can or cannot be infected with MPXV by infected owners?

What this shows is that pets are definitely being exposed from their owners (not surprising) and that it’s possible that some are getting infected. However, it’s unclear whether these were true infections, and even if they were true infections, the low viral load and lack of clear mpox lesions would suggest that the transmission risk posed by pets is probably very low to negligible.

To that end, we've a preprint today from researchers at the NIAID labs in Hamilton, MT that finds that the bovine (genotype B3.13) strain of HPAI H5 produces more severe neurological disease in C57BL/6J lab mice than did 3 other strains (two 2.3.4.4b isolates & a 20-year-old strain from Vietnam).

Curiously, tests on BALB/C mice produced more rapid onset of severe respiratory symptoms, but without the neurological involvement seen in C57BL/6J lab mice. While several plausible explanations were offered, the reasons remain unclear. 

-

In this study, we compared disease progression and pathology of three recent clade 2.3.4.4b isolates derived from a cow, mountain lion, and mink to a human HPAI A(H5N1) isolate from Vietnam in mice.

Inoculation of C57BL/6J and BALB/c mice with all four HPAI A(H5N1) isolates resulted in comparable levels of virus replication in the lung inducing severe respiratory disease.

C57BL/6J mice infected with the bovine isolate also developed high virus titers in the brain, resulting in a significant pro-inflammatory cytokine response and neurologic disease.

Our findings suggest the recent bovine isolate possesses enhanced respiratory and neuroinvasive/neurovirulent properties causing fatal respiratory and neurologic disease in C57BL/6J mice.

And as we've discussed previously, even minor genetic variations between viruses can produce vastly different characteristics (see  https://afludiary.blogspot.com/2012/04/differences-in-virulence-between.html - Differences In Virulence Between Closely Related H5N1 Strains ).  

While we may be comforted by the lack of severe H5N1 infection in the United States, the virus continues to rapidly evolve, making past performance no guarantee of future results.