test

Earlier today I wrote about a New York Times story about some interesting H5N1 flu transmission data from the US CDC that was posted briefly on Wednesday and was then quickly removed from the CDC website. The information suggested they’d identified cat-to-human and human-to-cat transmission of H5N1 flu in two separate households, both of which would be noteworthy events. A bit more information about the cases has since come to light, from outside official channels (the lack of transparency regarding these findings is a separate story).

Household 1

  • Cat 1 got sick, died, and tested positive for H5N1 influenza.
  • Cat 2 got sick 4 days after Cat 1 got sick. That’s pretty consistent timing for cat-to-cat transmission, but it appears Cat 2 was not tested for flu.
  • Cat 3 was tested for flu and was negative. Sorting out the testing for this cat is important, including when the cat was tested and what samples were collected, as this can impact how likely a the result is to be a true negative or a false negative.
  • An adolescent got sick 6 days after Cat 1 got sick / 2 days after Cat 2 got sick. The adolescent tested negative for H5N1 flu 6 days after the onset of illness. It’s not clear if there was testing for other diseases (I’d guess either that was not done or results were negative).
  • An adult dairy farm worker was in the household. That person was healthy and was not tested. There’s no mention of whether the cattle on the farm where the person worked were known to be infected with H5N1 flu (but I wonder if a farm exposure history is why Cat 1 was tested initially).

We still need more details to assess this further. The timing of disease in the cats and the adolescent certainly fit with H5N1 flu transmission within the household. The negative tests decrease the concern a bit, but sample quality and especially timing of testing can make a significant difference. False negative tests can occur when we sample too late in disease if the virus has been eliminated by the time we test (even if the individual has not fully recovered yet). That’s a particular concern for a virus like H5N1 flu (in its current form) that’s not well adapted to humans and other mammals (hopefully it stays that way); I’d expect longer viral shedding with a well-adapted virus. So in this case sampling later in the disease course doesn’t rule out H5N1 flu entirely. If the adolescent had flu-like disease and tested negative for other potential causes (e.g. SARS-CoV-2, human seasonal flu), then I’d still be concerned about H5N1 flu.

Based on the limited information available so far, I’d consider this to be potential but as yet unconfirmed H5N1 flu transmission from cat-to-cat and cat-to-human.

Knowing the strain of H5N1 flu in the cat will help too. If it’s the dairy strain but the cat had no direct contact with an affected farm (e.g. indoor only cat on a farm, or an off-farm cat owned by a farm worker), that would suggest that the cat got infected from its owner, either because the person had and asymptomatic / unrecognized infection or they tracked the virus home on their clothes. Hopefully they’ll do some sequencing, and serological testing too. Antibodies against H5N1 in either the cats or the people could tell us if infection occurred in the past, even if they weren’t sick. Positive serology in the remaining cats or the kids would be strongly supportive of household transmission.

Household 2

  • Cat 1 got sick, died a day later and tested positive for H5N1 influenza.
  • Cat 2 was healthy and tested negative for H5N1 influenza on the same day as Cat 1.
  • The owner was a dairy farm worker who got sick 2 days before Cat 1, but it appears the person was not tested (at least not for H5N1 influenza).
  • The household was lost to follow up 3 days later.

We can’t rule out transmission from Cat 1 to Cat 2 in this case because the time frame for both the initial test and the monitoring period were too short to say for sure that Cat 2 did not get infected, but there’s nothing to indicate cat-to-cat transmission here either.

The timing of illness in Cat 1 could fit with human-to-cat transmission, because people can be infectious before they start to feel sick, and/or have only mild signs of illness at the start that they don’t recognize as being flu-related. Only 2 days between the owner and the cat getting sick is on the short side, but still within the range we might anticipate for flu transmission with very close contact in a household.

Knowing the cats’ lifestyles and other risk factors is important too. If Cat 1 was indoor only and not fed raw meat or milk, the owner is the logical source of the virus. If sequencing shows the cat was infected with the dairy strain of H5N1 and the cat didn’t have access to dairy cattle or their environment, that would be a very solid link too. With the person being sick first, it’s strongly suggestive of human-to cat transmission, but the smoking gun (testing the person before the cat and finding the same strain in the cat and person) is missing.

What does this all tell us?

We still need more details. Presumably most or all of this information is known to someone involved in the investigation. Getting the information out is important to help others identify risks and keep messaging balanced, in order to avoid panic but also help prevent additional infections. Cat-to-human transmission concerns can raise a lot of fear (and potentially bad consequences for cats). Human-to-anything transmission would be really noteworthy, so we need to know if it was almost certainly occurred, maybe occurred or almost certainly didn’t occur.

I’ve already been working under the assumption that cats with H5N1 influenza can be infectious to others, and this doesn’t change that. It gives a little more support to the concern (and weight to the recommendations to help avoid it) but doesn’t answer the question completely. It definitely does not make me less concerned.

As usual with emerging diseases like this, we need more surveillance, including the details.