test

Last week, I wrote about the use of oseltamivir in dogs and cats infected with H5N1 influenza. We have to be aware of the risk of drug resistance risks whenever we use anti-infectives, especially when the same drugs (like osteltamivir) are used in people, and assess the risks and benefits in order to “use as little as possible, but use enough.”

That’s lead to some questions about use of the antiviral GS-441524 in cats. This drug has been a game changer in the management of a previously almost invariably fatal disease in cats called feline infectious peritonitis (FIP). 

When considering use of oseltamivir for treatment of H5N1 flu, there are a few things to our advantage:

  • It’s a narrow spectrum antiviral 
  • H5N1 infection and virus shedding are short term 
  • There’s no endemic H5N1 flu virus circulating in the cat and dog population
  • Cats and dogs are not common (if ever) sources of H5N1 transmission to people
  • We can implement infection control measures during the short treatment course required and estimated virus shedding period to contain the risk of spread of any resistant virus

The risk of a resistant H5N1 virus emerging during use of osteltamivir in a pet is therefore low, and the risk of any such virus spreading is even lower. While there’s some risk, with basic precautions, I think we can justify its use in infected pets under the right circumstances.

When considering use of GS-441524 (GS) for treatment of FIP, there are some very important differences:

  • GS is a broader spectrum antiviral
  • Enteric feline coronavirus (the virus that mutates to ultimately cause FIP in cats) can be shed by infected cats for months
  • Enteric feline coronavirus is a cat-adapted virus that can spreads very efficiently from cat-to-cat via fecal-oraltransmission

The risks of resistance when using GS in cats therefore differ according to the scenario.

1. Using GS to treat a cat with FIP

    • There’s a risk of emergence of resistant FIP virus within a treated cat. This would be bad news for the cat, but probably of limited broader risk since once enteric feline coronavirus becomes a cause of FIP, it’s not readily transmitted anymore. Odds are that the cat would not transmit the resistant virus further. We can’t say there’s no risk, but it’s low risk.
    • If the cat had concurrent intestinal infection with feline coronavirus, then there would be a risk of that virus becoming resistant and then spreading. One study reported fecal shedding of feline coronavirus in 61% of cats with FIP that were being treated with GS . Shedding dropped fairly quickly in most cats, which shows some likely impact of the drug, but it also shows that there’s some plausible risk of resistance emergence and transmission.
    • Since FIP is devastating, GS is highly effective, and the risk of resistance spreading is low, this is clearly a high-benefit / low-risk use situation. However, it’s not no risk so we need to study it more and optimize our treatment approaches.

    2. Using GS to treat cat with enteric feline coronavirus infection

    • Feline coronavirus is widespread and continually circulating in cats. There’s been some discussion of use of GS to knock that back, and to try to eliminate it from groups of cats (e.g. catteries). Treatment will reduce fecal shedding of the virus, and less shedding would likely have some downstream reduction of FIP, but I have my doubts that we can do much to control spread in the grand scheme with an antiviral. Reducing and eliminating a virus are different, and reducing while creating a substantial risk of resistance isn’t usually a good combination. In general, we are rarely able to use anti-infective drugs for effective infection control approaches in a population, especially for a virus that’s host adapted and endemic.
    • If we are treating cats with enteric infection, there’s a lot of virus, a lot of cats and a lot of chance for resistance emergence. If resistance emerged in a cat, it could shed large amounts of virus for long periods of time, releasing GS-resistant virus into the cat population and hampering our ability to treat FIP when it occurs. That’s a big concern for me.
    • Since enteric feline coronavirus infection isn’t a major health issue, treatment is not likely to have a major impact on enteric virus circulation, GS is so important for cats with FIP, and resistance would result in cat deaths, I have a hard time finding an indication for use of GS for enteric feline coronavirus.

    Dr. Niels Pedersen, a (or The) leader in development of antiviral approaches for FIP has a nice commentary entitled “Inappropriate use of GS-441524 in an attempt to eliminate Feline Enteric Coronavirus (FECV) from healthy cats.” The title gives away his thoughts on the matter. It’s a good, impassioned summary of why we need to be good stewards of FIP antivirals and why targeting feline enteric coronavirus is likely a bad idea.

    test

    For the longer answer to this question, check out the latest podcast episode I just posted on WormsAndGermsPod. For those who prefer to read the summary, here it is:

    If you ask people on the street “should we use the limited antivirals we have available to treat people with flu on dogs and cats?” the common answers would probably be “no,” along with the occasional “hell no!!” On the surface, that response makes sense. We have limited antivirals (both in number and sometimes supply) and they are important for treating people in some situations. However, we shouldn’t completely dismiss the utility of antivirals in animals with influenza infection in selected circumstances with good controls.

    Early treatment with an antiviral like oseltamivir (Tamiflu) could be effective against H5N1 flu in some animals, and there may be a rationale for prophylaxis in high risk situations (e.g. housemate of a cat with known H5N1 flu). Both of these scenarios involve companion animals, and non-reservoir species, and no livestock (including backyard chickens).

    Is there a concern about development of antiviral resistance if we use the drug in pets?

    We can’t ignore the possibility. Antiviral resistance is a concern, and it’s a spontaneous event that can occur as flu spreads, regardless of whether antivirals are used or not. Antiviral resistance markers have been found in influenza virus isolated from poultry (that were never treated with antivirals). The more the virus is transmitted and the more antivirals are used, the greater the risk of selecting for and spreading resistant strains.

    But antiviral resistance is different from antibacterial resistance in some important ways. Antivirals are much more specific in their effects, and viruses don’t swap resistance genes with each other like bacteria do.

    • If we treat an animal or person with an antibiotic, there will be myriad bacteria in that individual that are resistant, can become resistant or can spread resistance genes.
    • In contrast, if we treat an animal with oseltamivir when it doesn’t have the flu, it can’t select for a resistant flu strain.
    • If we treat and animal with oseltamivir when it does have the flu, there is a chance of resistance developing, but it’s only a broader problem if that flu virus is passed on to another individual. We don’t know if dogs and cats can transmit H5N1 flu, but it’s prudent to assume that they can. So, if we’re going to use an antiviral, we need to do all that we can to reduce the risk that any virus from that individual does not get transmitted to anyone else. We can probably do much more effectively in a pet than we can in a person that’s being treated with an antiviral.

    Good antimicrobial (including antiviral) stewardship means use as little as possible but use enough. We need to be prudent, but we also shouldn’t miss opportunities to intervene when we can do so effectively and with minimal risk.

    When does it make sense to use an antiviral to treat H5N1 flu in a dog or cat?

    Use of an antiviral makes sense for early treatment of known or high risk cases of H5N1 influenza where there’s a concern for development of serious disease (i.e. any infected cat, and probably infected dogs) AND when the animal can be properly treated AND when the animal can be kept isolated during and shortly after the treatment period.

    Basically my two main questions are: do they need it? and am I confident the animal won’t be able to infect another individual (human or animal)? If I can comfortably say yes to both of those, I think it’s reasonable to use an antiviral.

    Example 1: An infected cat in a household or veterinary clinic

    • Yes. We can properly treat, isolate, monitor and test the cat appropriately.

    Example 2: An infected cat that goes outside

    • No, unless the cat can be kept inside during the treatment and monitoring period. I don’t want to risk an antiviral-resistant flu strain developing and then the cat spreading it to other cats, or worse, birds.

    Example 3: A potentially infected backyard chicken

    • No (or hell no). These are livestock, so they are approached differently (and in Canada poultry infected with H5 flu must be culled). Poultry are highly susceptible to H5N1 influenza, and can clearly infect people. Also, an antiviral is probably too little, too late for a species that is so susceptible.

    Canine heartworm (Dirofilara immitis) is a nasty parasitic infection of dogs that’s relatively rare in Canada, but it’s still a concern because of how severe it can be and because treatment isn’t easy. It’s transmitted by mosquitos (we have lots of those), but transmission is also influenced by temperature (which varies a lot in Canada over the year).

    Heartworm control focuses on regular administration of preventive medications that are typically given to dogs monthly, either during the heartworm transmission season or all year round. The medication is often combined with flea/tick preventives and/or other antiparasitics.

    We have a seasonal risk period for heartworm in Canada, and it’s pretty short in some areas. We want to maximize use of heartworm preventive drugs during that period, so understanding exactly when the risk is present is important. Since transmission is temperature-dependent, the transmission period will vary across the country, and as out climate changes over the years, we need to keep re-evaluating the timing for giving dogs preventatives.

    Reminder: Heartworm life cycle

    Canids (including dogs and wild canids, like coyotes) are the reservoirs for this parasite . Adult heartworms produce microfilaria which are found in the dog’s bloodstream. These are picked up by mosquitoes when they feed on a dog. They then develop into the mosquito from microfilaria to L1, L2 and then L3 larvae (L3s are the infectious form). At that point, if the mosquito bites another dog, it can transmit heartworm via the L3 larvae.

    Why is temperature important?

    Heartworm larval maturation stops at temperatures less than 14C (57F), so the time that it takes for an infected mosquito to be able to transmit infectious heartworm larvae depends on the time that’s above 14C, and how far above 14C the temperature is. We measure this in “heartworm development units” (HDUs), which are calculated by taking the mean (not maximum) daily temperature in Celcius and subtracting 14. If it’s a positive number, that’s the number of HDUs for that day.

    Today, the mean temperature will be around 0C here. So, if there was an infected mosquito, the heartworm larvae are not developing. Hopefully things will warm up soon. If we get a day when the mean temperature is 17C, then we’ve accumulated 3 HDUs for that day. Once we get a total of 130 HDUs, those larvae should be infectious. That obviously takes time in spring in Canada, but not much time in places like the southern US when mean daily temperature is consistently and substantially higher, and 130 HDUs can be accumulated quickly, any time of the year.

    In Canada, June 1 has traditionally been the recommended starting date for heartworm prevention. That was based on temperature data from Windsor, Ontario (pretty much as far south as you can get in Canada) plus a very safe buffer period. However, since data for this recommendation were from 1957-1986, we figured it was worth rechecking, and we published the results in December in the Canadian Veterinary Journal (CVJ)(Weese & Peregrine, 2024).

    The short answer: our current approach is still good.

    We looked at temperature data from 1996-2023 for a variety of cities in Ontario, and how long it took for 130 HDUs to be accumulated. (I’ve put together data for cities in other provinces too but haven’t done anything with those yet.) The results are shown in the table below. In southern Ontario, the risk typically started mid June. In Windsor, it was as early as May 25, but with a median of June 7. Some years, the risk started quite late, and in some northern cities it didn’t start until August.

    Okay, but May 25 is before June 1, so is it a problem to still use that date as the start of the risk period, at least in Windsor?

    • No. Fortunately, our heartworm preventive drugs will kill L3 and early L4 stages, so they will kill larvae up to 4 weeks after infection. If a dog was infected May 25, preventive administration as late as June 22 would still work.

    If we want to look at the limits for when we should start preventive treatment, it’s actually the date that 130 HDUs are accumulated plus 28 days. I definitely wouldn’t want to cut it too close, and since we don’t track HDU accumulation in real time (and there are some potential disclaimers), we want to give ourselves a good cushion in terms of timing.

    What about climate change?

    Climate change is real, but mean temperature changes are slow and gradual. The graph below shows the trends over time. In some locations, there was a significant trend towards earlier dates of 130 HDU accumulation. In others, there was actually a trend towards later. Overall, there was no significant change. However, it’s fair to expect that there will be gradual changes over time, so we’ll just need to keep checking in.

    Can we trust the 130 HDU requirement?

    Mother Nature’s always a bit unpredictable, but the 130 HDU threshold has been validated pretty well. Even if the threshold was lowered to something like 125 HDUs, it would only change the typical time to accumulation by one or two days. 

    When we’re assessing risk periods, we want to err on the side of over-estimating, not under-estimating risk, so we took a conservative approach with our methods. You can calculate HDU accumulation two ways: 1) simply calculating the time until 130 HDUs is hit in the spring, or 2) using a 30 day rolling window. The latter is based on an assumption that adult mosquitoes only live for a maximum of 30 days, so early warm periods become irrelevant if the mosquito dies before larvae can become infectious. We chose to use the total time, not the 30 day window, since there is evidence that some mosquitoes can live for over 30 days.

    There are also potential issues such as microclimates (mosquitoes living in a warmer local site like a building or sewer) and heat sinks. Whether this is relevant to heartworm transmission is unclear, but it’s probably a minor issue, if it’s an issue at all.

    What does this mean for routine heartworm prevention?

    It means that it’s still fine to use June 1 as the start of the risk period in Canada, even in extreme southwestern Ontario. We don’t need to worry about starting earlier (a question I get all the time), at least not yet. June 1 still provides a good cushion too, for situations where people are late getting or starting the prescribed drug.

    But… don’t forget about ticks!

    Ticks are different. The tick risk season is much longer than the heartworm risk season. We have some risk of tick activity anytime the temperature is above 4C (or maybe even lower). Obviously, that means we have risk of tick exposure well before risk of heartworm exposure, and we have problems with people starting tick prevention too late because they only think about the June 1 heartworm risk date. If ticks are a concern, then tick prevention earlier in the year is important (and if it’s combined with heartworm preventive, the heartworm start date discussion becomes a moot point).

    test

    When it comes to food safety risks with H5N1 influenza, we know it’s a problem with raw (unpasteurized) milk from cattle, so I often get asked if it’s a concern with milk from other animals, such as goat’s milk. It’s a good question. My typical response has been that we really don’t know. There haven’t been any reports of H5N1 flu in other dairy species (except for one report at this time last year of infection in neonatal goats on a farm in Minnesota that had infected poultry), but we have a lot more dairy cattle living on much bigger farms in North America. So it’s hard to say if cattle are actually at greater risk, or if we’re just more likely to recognize problems on a big dairy where the animals are monitored closely.

    There is now a report from the UK about H5N1 infection in a single ewe (a female sheep) in Yorkshire, which was on a farm that also had infected captive birds. While this doesn’t tell us about the risk in goats or the overall difference in risk between dairy cattle and other dairy animals, it shows that cattle are not the only dairy species of concern. 

    The ewe tested positive on a milk sample, which was collected as part of the response to the infected birds on the farm (a smart surveillance approach that helps with early detection of spillovers and helps us understand transmission risks and patterns). Only one sheep from the flock was positive, but we have to assume the risk is likely broader – if one sheep can be infected, then other sheep could be too, if exposed to enough of the right virus.

    The infected sheep was culled, which is understandable, as it removes the risk from that sheep. Unfortunately but it also removes any ability for us to learn more about what H5N1 flu does in a naturally infected sheep. It would have been ideal to isolate the sheep and monitor it for signs of illness and test it for virus shedding, but that’s not always possible. 

    There are a few good take-aways from this report:

    • It shows the value of surveillance. If they hadn’t tested the other animals on the farm, we’d have no idea the sheep was infected, as the sheep did not appear sick.
    • Presumably, this was a direct spillover from the infected birds on the property. Hopefully that means the virus in not established in sheep anywhere else (like it is in dairy cattle in the US), and that this was an interesting one-off infection but not of broader concern.
    • If there really was only one infected sheep, removal of that ewe may have prevented it from infecting other sheep, and the other sheep all tested negative. I hope there will be a bit more testing on this farm to make sure there are no secondary cases.

    I guess we can’t rule out the possibility that the sheep infected the birds. However, since they tested all the other livestock and only this one sheep was infected, it’s pretty safe to assume this was a direct bird-to-sheep spillover. Sequencing of the virus will help confirm this. Serological testing of other sheep on the farm would be interesting to see if there’s evidence of earlier infections.

    Does finding H5N1 flu in a single sheep change anything?

    No, not really. We know that H5N1 has spilled into a wide range of different mammals, and this just expands that list. We’ve assumed there’s some risk from other dairy livestock species, and this shows that’s a reasonable concern. We’ve also talked about risks from raw milk even before H5N1 was concern, there are lots of infectious disease risks with drinking any type of raw milk.

    However, this report does raise the stakes a bit. We don’t want continued spillovers into mammals, because that increases the risk of this virus adapting to become better able to infect more mammals (including people). We don’t want endemic transmission in domestic mammals, as that increases human exposure risk. We also don’t want H5N1 flu in the food supply (but remember that pasteurization will kill the virus). 

    UK Chief Veterinary Officer Christine Middlemiss’ statement sums things up well:

    “We have confirmed the detection of influenza of avian origin (H5N1) in a single sheep on a farm in Yorkshire. Strict biosecurity measures have been implemented to prevent the further spread of disease.  

    “While the risk to livestock remains low, I urge all animal owners to ensure scrupulous cleanliness is in place and to report any signs of infection to the Animal Plant Health Agency immediately.” 

    The UK Health Security Agency (UKHSA) has said that avian influenza is primarily a disease of birds and the risk to the general public’s health is very low, but people should not touch any dead or sick wild birds they find. 

    The Food Standards Agency advises that properly cooked poultry and poultry products, including eggs, remain safe to eat and avian influenza poses a very low food safety risk to UK consumers since the H5N1 virus is not normally transmitted through food.”

    Awareness.

    Diligence.

    Good use of routine infection control and biosecurity practices.

    Continued surveillance.

    Those are the key factors for limiting the risk from this virus. 

    test

    It’s fair to say that raw pet foods have clearly been source of H5N1 influenza infections in cats, some of which have been fatal – even though sometimes it’s impossible to make a definitive link between the food and infection because of when and how the infection occurred, and what could or could not be tested in timely manner. Sometimes raw pet food companies and their supporters use that missing link as a way to claim that there’s no risk or that the risk is overblown. Confusing (and let’s just say “opportunistic”) communication around test results also adds to consumer confusion – for example, understanding the meaning of “non-negative” test results.

    The two most common ways in which the term “non-negative” is used with regard to test results are:

    1. Borderline positive: The test result was not quite strong enough to say it’s positive, but there’s a strong enough signal to make us very suspicious that there’s something there (i.e. the virus of interest). Since we can’t call it a positive or a negative, “non-negative” indicates that the sample could be positive, but we can’t say for sure.
    2. Reportable disease that requires government laboratory confirmation: For some diseases (usually really important ones, like reportable diseases), only the government lab gets to have the final say on the test result. If another diagnostic lab gets a “positive” result (even if it’s a really clear positive), they’re only allowed to call it “non-negative” until the result is confirmed by the government lab. That might mean running the same test, a similar test, or a complementary test. Sometimes, that confirmatory test might actually be less sensitive (i.e. less able to detect a positive), but is nonetheless the test used to provide the definitive answer to determine what response is required.

    When we were doing our SARS-CoV-2 surveillance of pets in households with people with COVID-19, we had a few results that were positive/non-negative in our lab but were very close to the cut-off level, and were ultimately just below the cut-off at the federal lab. Those had to be called negative, even though we were pretty confident the virus was present in those animals, such as when we had a clearly positive dog (confirmed by the federal lab) and a non-negative dog in the same household. Odds are very high that the second dog with the borderline result was just shedding less virus than the first dog. But I digress…

    Savage Cat Food has issued a recall of some of its raw chicken diets (lot code/best by date 11152026) following infections that were linked to the food and “non-negative” test results for H5N1 flu. . The food was distributed to California, Colorado, New York, Pennsylvania and Washington (not Canada).

    • In February, the company was made aware of a cat in Colorado that developed H5N1 flu and that had been fed their food. The cat fortunately recovered.
    • Sealed packages of the food were sent to the Colorado State University Laboratory and were tested “non-negative” by PCR for H5N1 flu.
    • A sample was then sent to the National Veterinary Services Laboratory in Ames, Iowa for virus isolation testing, which was negative. However, virus isolation is less sensitive than PCR, so PCR positive/virus isolation negative results can definitely occur with a contaminated sample, especially as flu virus is likely to die over time sitting in the food (virus isolation requires the virus to still be viable, whereas PCR can detect “dead” virus).
    • While this was being done, the manufacturer contacted retailers to have them pull the product from shelves.
    • A week after the negative virus isolation result was obtained, the company got a report of another cat with H5N1 flu that had eaten the same lot of food – in New York.

    Some people will point to this and say “there’s no definitive proof” that the cats were infected by eating the food. Yes, that’s true, it’s not definitive, but it’s still solid. If they are able to sequence virus from both cats (meaning they find the same virus in two cats from two different states that were fed the same diet that had a PCR positive result) that makes it an even more solid presumptive link. If they are testing other food samples and get a positive, or even just more non-negatives, it reinforces the link further. If they are able to sequence the same virus from the food as they found in the cats, that would essentially be a slam-dunk.

    The pet food company has a pretty straightforward influenza alert link on the home page of their website. They’re not saying the food was contaminated, since they indicate it tested negative at the Ames lab, but they are also (unlike some others) not trying to downplay the risk or deflect, so I give them credit for their response. (That said, I still take issue with various statements in their their general FAQ about food safety.)

    This company does not high pressure pasteurize their diets. On with website, they say “No, our cat food is not subjected to High Pressure Processing (HPP). HPP can damage helpful bacteria and can change the taste and texture of foods, often times making it unappealing for cats.” That’s likely a big risk factor for why the food is contaminated and why these cats got sick. High pressure pasteurization isn’t perfect, but it’s a good tool to reduce contamination and (despite what the website says) there’s no real argument not to high pressure pasteurize raw pet food, especially high risk poultry-based diets.

    The response of raw pet food companies to H5N1 influenza has been really variable.

    • Some have taken it seriously, talked about their risk reduction plans, and not tried to deflect.
    • At least one has started cooking their poultry diets.
    • Some have deflected and tried to downplay any risks.
    • Some are not saying anything.

    Hopefully more of these companies are exploring high pressure pasteurization as a risk reduction (though still not risk elimination) method for customers who are still intent on feeding raw diets to their pets.

    test

    A recent opinion article in the Washington Post about how to protect cats from H5N1 influenza includes some good considerations, but alongside some bad information. It’s prompted me to write a bit about pet food preparation and how it does (and does not) affect the presence of infectious pathogens (but please keep in mind while reading this that I’m an infectious disease veterinarian and not a food safety specialist!).

    In the article, there’s a statement from a professor at a school of public health who said they were avoiding canned pet food “…largely because the pet food industry is not as closely regulated as the human food industry,” …Coleman has not been giving her cats canned food, as it might contain unpasteurized milk. Kibble, on the other hand, is heated at very high temperatures, which, like pasteurization, would kill H5N1.”

    However, canned pet food is actually likely the lowest-risk form of pet food when it comes to pathogens. Processing conditions are very different for different forms of pet food, and the processing significantly impacts pathogen survival:

    Canned pet food

    These diets are heated to high temperatures (235F/115C or above), akin to sterilization. If it wasn’t, cans would spoil quickly and often explode because canned food is otherwise a great culture medium for bacteria. Canning will absolutely kill H5N1, full stop. No question, no concerns. Move on please.

    Kibble / dry pet food

    These diets are also thoroughly cooked prior to being made into kibble. The temperatures used are not standardized, but are well above the recommended temperature for cooking poultry (165F/74C). Heat is also generated during extrusion process (forming the food into kibble) and desiccation (being dry food) will also decrease pathogen survival. The food isn’t sterile, but as long as the finished product is kept away from raw ingredients (a pretty standard food safety step), I have no concerns about flu contamination in dry kibble.

    Lightly or gently cooked diets

    These diets are in the grey zone. They’re marketed as diets that are cooked but are not to the same extent as kibble. Killing pathogens in food is a function of temperature and time. If they reach the recommended cooking temperature for poultry (165F/74C), any flu virus in the product should be long gone. At lower temperatures, there’s more variability and more need for a longer cooking time (e.g. 60C/140F for 30 minutes). A challenge with these diets is the lack of information about the “gentle” cooking parameters (temperature and time). I suspect they likely all hit a high enough temperature to at least kill flu viruses (which are easier to kill than some food borne bacteria) but that’s a bit of a guess. I’d consider them low risk overall, but if I was going to feed a diet like this I’d ask the company how they cook the food, to be sure.

    High pressure pasteurized raw diets

    High pressure pasteurization (HPP) uses pressure (instead of heat) to inactivate microorganisms. However, the effectiveness is dependent on the amount of pressure and how long it’s maintained, as well as the food matrix itself, and volume of food undergoing HPP. If the method is validated and performed properly, the risk of flu virus surviving the process should be really low. However, we’ve seen at least one report of fatal H5N1 influenza in a cat that was eating a HPP-treated diet. Recalls of HPP treated diets for Salmonella contamination are also far from rare (and if the process doesn’t kill Salmonella in a certain product, we’d be concerned it might not effectively kill flu virus either). I consider HPP a risk reduction method, not a risk elimination method. If a company can show that their method kills flu (or a proxy virus), then I’d be more confident the diet is safe (at least from a flu standpoint). 

    Raw diets with “natural preservatives”

    Preservatives are used to reduce spoilage. That means they reduce growth of bacteria that are already in the food. They are not designed to kill pathogens (bacteria or viruses) in the food. So, whether or not there are preservatives (and whether or not they are considered “natural”) likely has no impact on survival of and risk of contamination with H5N1 flu in the diet.

    Freeze-dried pet food

    These diets are frozen, and then the water is removed under vacuum. Freeze drying is actually great method for preserving viruses for long-term storage in laboratories. I’m not aware of any evidence of any impact of freeze drying on survival of flu viruses, so until then I would assume that freeze dried diets are the same risk as fresh diets.

    Raw (fresh) poultry

    If H5N1 flu is present in a bird, it will be present in the meat from that bird. If there are no steps to kill it, flu virus can persist for a while in that nice moist environment, even after the bird has been slaughtered or died. Refrigeration probably helps with survival of the virus. Freezing probably has a bit of impact, since freeze-thaw cycles can impact viral viability, but I have no confidence in simply freezing the meat to significantly reduce viral contamination, and would not consider frozen raw diets to be any different than fresh in terms of risk. Untreated raw poultry is undoubtedly the highest risk pet food when it comes to H5N1 flu (not to mention a lot of other pathogens as well).

    Over a year ago, I wrote a few posts about changes to Clinical and Laboratory Standards Institute (CLSI) veterinary antimicrobial susceptibility testing guidelines. When a bacterium is grown in a lab and tested for susceptibility to different drugs, the lab looks to these CLSI guidelines to determine whether they should consider the bug susceptible or resistant to each drug. That’s critical information for choosing an effective drug to treat a patient with an infection.

    In January 2024, some major changes were made to these guidelines, particularly with reporting susceptibility to enrofloxacin and marbofloxacin in Staphylococcus, E. coli (and related Enterobacterales) and Pseudomonas, as well as chloramphenicol susceptibility in Staphylococcus and E. coli/Enterobacterales from dogs.

    I was hoping veterinary diagnostic labs would be able to implement those changes fairly quickly. I had a lot of empathy for labs given the challenges of sourcing different testing panels and making a big change to their methods and IT workflow. Some labs made the change fairly efficiently. However, the labs that likely account for the majority of veterinary diagnostic testing in North America still haven’t made the switch.

    I’m still empathetic to their challenges, but I’m seeing too many treatment failures because veterinarians are using drugs that the lab reported should work, when in reality the updated guidelines tell us the won’t (or we can’t interpret the test results properly because of what they tested or reported). That’s a problem, and I have no doubt that animals have died because of it.

    So, in addition to my detailed posts from last year about the CLSI breakpoint changes, here’s a (hopefully) quick and easy reference for interpretation of bacterial culture and susceptibility test results:

    test

    I get variations of this question not uncommonly from veterinarians:

    A client just came in and said they/their kid was diagnosed with Dientamoeba fragilis infection and their physician said to test/treat their dog. What should I do?

    On one hand, it’s nice to see physicians thinking about animal exposures. On the other hand, it sometimes ends up with the physician making a testing or treatment suggestion (or directive) for an animal that doesn’t make sense (or at least needs to be considered with a lot more nuance).

    Here’s the basic need-to-know about D. fragilis:

    • It’s a microscopic, protozoal parasite that lives in the large intestine. It’s shed in feces, and people get infected through fecal-oral exposure (aka eating poop, albeit in minute quantities).
    • It’s found in humans worldwide, and people are its normal host. Compared to many other intestinal parasites, it’s actually fairly common in people even in high-income countries. For example, a study of children in Toronto daycares found that 8.6% were shedding the parasite, along with 4.0% of the daycare workers (Keystone et al. 1984).
    • Many, if not most, people carrying this parasite in their intestine are healthy. It’s role in causing disease (if any) isn’t clear. If it does cause any illness, it’s pretty non-specific things like diarrhea, abdominal pain, nausea, fatigue and loss of appetite.
    • It’s been found in a few animal species, including non-human primates (e.g. gorillas), budgies and pigs. However, it seems rare and data are quite limited.

    One study (Chan et al. 2016) reported finding D. fragilis using PCR in a single dog and a single cat . Whether the dog and cat were truly infected or the test just cross-reacted with a different protozoan isn’t clear. If it was true infection, the low prevalence in dogs/cats compared to humans and the lack of evidence that infected dogs/cats have much of a parasite burden make it pretty unlikely that dogs/cats are a significant source of human infection. The aforementioned daycare study from Toronto (Keystone et al. 1984) reported an association between the parasite and cat ownership, but it’s far from certain that there was any causal relationship; the analysis was pretty basic and the relationship between cat ownership and other identified risk factors wasn’t explored. (The authors indicated that the association could have even occurred by chance.)

    I would consider D. fragilis a parasite that’s not normally found in dogs and cats, and is not well adapted to them if they are exposed. If a parasite infects people and is shed in feces, there’s a reasonable chance dogs and cats would be exposed (one way or another), but not necessarily that they’d get infected. Even if infection did happen, it could very well be at such a low level that transmission wouldn’t occur (i.e. a dead-end infection).

    At most, dogs and cats are rarely infected with D. fragilis, and there’s no evidence that they’re infectious to people. I never say never with diseases that we haven’t studied much but, it clearly spreads quite commonly human-to-human, so pets are not likely to play a relevant role in the transmission cycle. If anything, the risk is human-to-pet transmission, not the other way around. 

    Can we test pets for D. fragilis, if requested?

    Routine fecal flotation would detect this parasite if there was a reasonable burden. PCR could also be used, but there are no commercially available tests for animals (although a human PCR test would presumably work – but few human labs will accept diagnostic samples from animals).

    Should we test pets for D. fragilis, if requested?

    The key question for me with any test is, “what would I do with a positive vs a negative result?”

    If a pet tests positive for D. fragilis, I’d say “Interesting. Your dog presumably got this from you or another person with whom you have contact. Odds are really low that the dog will infect anyone. It’s a don’t-eat-poop disease, so let’s just focus on basic hygiene and safe fecal handling, and it will go away on its own.“

    If a pet tests negative for D. fragilis, I’d say “It’s probably negative, but testing isn’t 100% sensitive. Regardless, we don’t want you to give it to your pet and we don’t want your pet to give you any of the many bacteria and parasites that might be in its intestine, so let’s just focus on basic hygiene and safe fecal handling.”

    Not a lot of difference either way.

    Should we just treat pets for D. fragilis is someone in the household has it? 

    If I had to treat a D. fragilis infection in a dog – where we found the parasite AND had reason to believe it was causing disease (both pretty unlikely) – I’d most likely use metronidazole. If the parasite isn’t causing disease I’d want to avoid using an antibiotic like that, because it’s not an innocuous drug and it will negatively impact the dog’s normal intestinal microbiota, potentially for weeks.

    Time is the cheapest, safest and ultimately most effective treatment in a healthy dog that happens to be infected with this parasite. More than that though, if people in the house are infected, they’re better off focusing on their hygiene (i.e. not eating poop) and eliminating their own infection, rather than worrying about the pets.

    test

    We don’t talk a lot about antiviral resistance in animals, particularly compared to antibacterial resistance, primarily because we don’t use antivirals a lot in animals – but we do use some.

    When we use any anti-infective medication, we have to think about the risks of resistance developing to that medication, and how we can try to optimize use in order to maximize clinical benefits while minimizing resistance risks. I say “minimizing,” not “eliminating,” on purpose, because we can rarely guarantee that there will be no risk. Usually we’re aiming for a risk that is low enough to justify use of the drug, and ideally we can lower that risk even further through some basic infection control practices.

    I’ve written a lot over the last year about antivirals for the treatment of feline infectious peritonitis (FIP) in cats, since we now have incredibly effective options (now in Canada) for what was previously an almost invariably fatal disease. The antiviral drug GS-441524 (a relative of the COVID-19 antiviral drug remdesivir) is what we can now use in most regions to treat this disease. It’s a complete game changer.

    However, anytime we have a revolutionary new anti-infective (be it for bacteria, viruses, parasites, fungi or other nasties), we need to take steps to try to maintain its effectiveness long term – primarily through good antimicrobial stewardship.  

    Are there scenarios where GS-441524 resistance could emerge in cats and be a problem? Yes.

    Does our routine treatment of FIP pose much of a risk for emergence of antiviral resistance? No.

    Those statements seem to conflict, but let me explain:

    Feline infectious peritonitis is caused by feline coronavirus (FCoV), whichbis very common in cats, especially especially cats living in large groups. It infects the intestinal tract and then cats shed it in their feces, usually with no signs of illness The virus is spread cat-to-cat through fecal-oral exposure. The clinical disease (FIP) develops when an intestinal-dwelling FCoV mutates into a tissue-infecting FIP virus (FIPV) within that cat.

    While FCoV is highly transmissible, FIPV is not traditionally considered to be transmissible between cats (though I’m not sure we can actually say there’s no risk of transmission; I think there is some cat-to-cat transmission risk, but it’s probably very low).

    Antiviral resistance can emerge when a virus is exposed to an antiviral. Random mutations occur that allow viruses to resistant the action of the drug, and then the mutated virus multiplies while the susceptible version may die out. The key is what happens next, and that varies a lot depending on the disease.

    A critical component of the risk assessment for GS resistance is that is that we don’t consider FIPV to be transmissible (or at least we consider it really rarely transmissible). So if GS resistance develops in a cat with FIP, while it could be very bad for that cat (because the infection won’t respond to the drug), the odds of that cat transmitting the resistant virus to another cat are very low, so the resistant virus would likely die out with the first host.

    The main concern is if cats with intestinal FCoV infections are treated with GS, then resistance could develop in FCoV in the gut of a cat, and that resistant virus would be shed in feces and be transmitted cat-to-cat (like FCoV normally is) through the cat population. Then, if any of those cats develops FIP, the virus would already have GS resistance and wouldn’t respond to treatment.

    That means that appropriate treatment of cats with FIP poses an exceptionally low risk of causing resistance issues. For there to be a significant problem, it would require rare cat-to-cat transmission of FIPV (or emergence of resistance of FCoV from a concurrent, unrelated FCoV infection in the intestine). The benefits of treatment massively outweigh this small theoretical risk.

    Empirical treatment of cats for FIP (without a solid diagnosis) is also generally low risk of resistance. If the cat doesn’t have FIPV or FCoV, there’s no risk. It’s not like the situation with bacteria where there’s always a massive and diverse pool of bacteria in the body that could develop resistance through exposure as “bystanders” and where resistance genes can then be transmitted between different bacteria. If the cat doesn’t have FCoV or FIPV to start, it can’t develop resistant FCoV or FIPV. The only risk would be (as above) if there was a concurrent incidental intestinal FCoV infection.

    Does use of GS in cats with enteric FCoV pose a risk of resistance? That’s a big “hell yeah!” from me. We don’t have data to prove it, but the last thing I want is to be treating lots of cats with highly transmissible enteric FCoV infections, because spreading resistant FCoV through the cat population is how we end up with a lot of dead cats from resistant FIP.  The cost-benefit just doesn’t add up in this scenario. Enteric infections are mild or cause no disease at all, and we are not going to eradicate FCoV from the overall cat population, or likely even from a local cat population (e.g. cattery), using antivirals (the virus is just too common and transmissible to be thinking about eradication). So we’re talking about treatment with little to no clinical benefit and no eradication benefit, while posing significant risk of selecting for resistance to a game-changing drug that is currently saving countless feline lives. That makes no sense to me.

    Are there any human health issues related to use of GS in cats?

    GS is related to remdesivir, a drug used to treat COVID-19 in people, but transmission of antiviral resistance from FCoV to SARS-CoV-2 isn’t a risk. The only theoretical risk would be if a cat had FIP AND a SARS-CoV-2 infection at the same time, AND resistant SARS-CoV-2 emerges in the cat during treatment AND the cat then infects a person with it. That’s a really unlikely scenario, so the risk is negligible compared to risks from direct use of remdesivir in humans.

    Use as little as possible but use enough. That’s my line for antibiotics, and it applies equally well to antivirals. We don’t want to miss treatment opportunities, but can’t be reckless either and jeopardize this incredible new opportunity to be able to treat this terrible disease.

    Details have now been released about the information that was briefly posted then removed from the US CDC website a couple weeks ago regarding suspected household transmission of H5N1 flu to and from cats. It’s great to have more details, but (as is common with emerging diseases) the full story probably raises more questions than answers. The gist of the story is similar to what we inferred from the leaked information, but the full report is now available online (Naraharisetti et al. MMWR 2025), and includes some useful details. Here’s the more complete rundown:

    Household 1

    • This household contained 2 adults, 2 kids and 3 indoor cats. One of the owners worked on a dairy farm that was not known to have infected cattle, but there were infected farms in the area.
    • Cat 1: One of the adult indoor cats developed loss of appetite, lethargy and disorientation, then worsening neurological signs.  After a few days, she was euthanized because of rapid and severe disease progression. H5N1 influenza virus (genotype B3.13) was detected in the cat’s nasal passages and brain tissue; this is the predominant strain currently circulating in US dairy cattle. 
    • Cat 2: One of the other cats then developed signs of respiratory disease four days after the first cat got sick – that timing is generally consistent with what we’d expect for transmission of a flu virus. The owner was asked to collect a nasal swab from the cat because it was too sick to be taken to a veterinary clinic, but this wasn’t done. Fortunately, the cat recovered after 11 days.
    • Cat 3: The last cat just watched everything unfold and appeared healthy the whole time.

    The household investigation:

    The owner who worked on the dairy farm didn’t have direct contact with cattle. He reported that he took off his farm clothes and boots when he returned home and that they were stored in an area that the cats could not access. The cats were not fed a raw diet or raw milk.

    One of the kids had mild upper respiratory tract disease (cough, sore throat, headache and muscle pain) six days after the cat got sick. The other kid had a dry cough around the same time that was attributed to allergies. 

    Eleven days after Cat 1 got sick, samples were collected from three of the people in the household. That’s a bit later than ideal, but still reasonable (logistics often dictate when samples can be collected). All the people tested were negative for flu. The sicker kid had a positive test result for rhinovirus/enterovirus, which could explain the illness, but it doesn’t rule out flu completely. 

    Unfortunately, the person who worked on the dairy farm declined to submit a sample. He reported a day of vomiting and diarrhea before Cat 1 got sick… Was that potentially related to flu infection? Maybe, since flu can cause gastrointestinal disease, but it’s far from classic, and most that episode was unrelated. The unanswered question is whether the person had a mild or asymptomatic H5N1 flu infection and was responsible for infecting the cat(s).

    The people who were tested for flu also got oseltamivir prophylaxis (antiviral medication) because of their close contact with the infected cat (and potentially the initial source of the virus in the household…); the farm worker declined prophylaxis.

    Putting Household 1’s story together:

    • The most likely source of H5N1 flu exposure of Cat 1 was the farm worker. Whether he unknowingly tracked the virus home from the farm or whether he had a mild infection that he transmitted to the cat can’t be discerned, but that’s a really important thing to figure out in future cases.
    • It would also be really useful to know if the farm where the person worked had confirmed H5N1 flu in the cattle. Hopefully follow up testing of the cattle was done. If flu wasn’t present, that raises more concerns about how the farm worker might have gotten infected, or if the farm worker was not the initial source in the household, then how did an indoor cat with no other apparent risk factors get infected? 
    • It’s unclear whether there was cat-to-cat transmission of flu. Spontaneous, transient upper respiratory tract disease is uncommon in adult indoor cats. Maybe stress caused a recrudescence of feline herpesvirus infection in Cat 2, but it sounds like the cat was a lot sicker than we’d expect for that. We therefore have to consider Cat 2 to be a strong flu suspect; based on the timing of illness relative to Cat 1’s illness, if it had flu, it presumably got it from Cat 1.

    Ideally, serological testing would be performed to detect anti-H5N1 antibodies in blood from individuals (cats and people) in the household, to determine which individuals ultimately were infected, even though serology can’t tell us the when or how.

    • If the farm worker was seropositive, it still wouldn’t tell us if they infected the cat directly or if the cat was exposed to virus on clothing or other fomites (though the former seems more plausible, and would be concerning). If the farm worker was seronegative, that suggests that clothing or fomites was the source (which is a good reminder of why infection control measures like wearing and changing personal protective equipment are so important).
    • If Cat 2 was seropositive, that doesn’t confirm cat-to-cat transmission, though that would still be most likely.
    • If the kids or other adult were seropositive, then there was almost certainly either cat-to-human or human-to-human transmission, neither of which would be good news.

    Household 2

    This household consisted of a dairy worker and two indoor cats. It’s not clear whether the person had contact with cattle, but he collected raw milk from farms and reported being frequently splashed with milk. He didn’t remove his work clothing before returning home, and one of his cats liked to roll on his work clothes (that were presumably contaminated with milk).

    • Cat 1: The worker’s six-month-old indoor cat was taken to a referral veterinary hospital with a one-day history of progressive neurological disease. It was seriously affected and died in hospital within 24h. H5N1 influenza virus (genotype B3.13, the dairy-associated type) was confirmed in this cat. Whether it was a coincidence or not, this was the cat that liked to roll in the work clothes.
    • Cat 2: The other cat remained healthy, and swabs collected from this cat the day after Cat 1 got sick were negative for flu. 
    • The worker reported signs of eye irritation that began 2 days before Cat 1 got sick. The person was not tested and declined antiviral treatment. So we don’t know if the person had H5N1 flu, but eye inflammation is consistent with the conjunctivitis seen in other human cases of H5N1 infection in the US in the last year, and would be a likely presentation in someone getting splashed in the eyes with influenza-contaminated milk.

    There was little further follow up because the owner shut down contact with public health officials, stating a fear of losing his job. 

    As with Household 1, timing of sample collection is always a challenge. While we’d ideally get serial samples over time, it’s simply not possible in many cases. Testing Cat 2 at the start was logical and useful, but if there was cat-to-cat transmission, that was probably too early to test. We’d need at least one more sample a few days later to account for the incubation period if Cat 1 infected Cat 2. The initial sample only tells us that Cat 2 likely wasn’t exposed at the same time as Cat 1.

    It’s also hard to say if the owner had H5N1 flu or if his clothing was the source of virus for Cat 1. Ideally they could have done serological testing after the fact to confirm whether or not the person was infected, but it doesn’t sound like the person would have been willing to come back to provide a blood sample.

    What about the veterinary staff involved in these cases?

    Twenty-four (24) veterinary clinic personnel were potentially exposed to the two confirmed H5N1 flu-infected cats from the two households, 18 of which were contacted and monitored for signs of flu. They were all deemed to have had limited exposure. Hopefully that’s a win for routine infection control practices and quick identification of higher-risk situations warranting use of enhanced personal protective equipment (PPE). Because they were all deemed low risk, they were not offered antiviral prophylaxis. Seven individuals reported signs or symptoms of illness after exposure, of which five were tested for flu, and all were negative.  

    To sum up:

    • Was there direct human-to-cat transmission of H5N1 influenza? Maybe.
    • Was there direct cat-to-cat transmission of H5N1 infleunza? Maybe.
    • Was there transmission of H5N1 influenza via contaminated clothing? Maybe.

    Unfortunately, we need to learn more about transmission by investigating disease events. It’s almost certain there are undiagnosed infections, and maximizing our recognition of H5N1 flu spillovers between species (including humans) is critical, as it helps inform proper care and management of people and animals (including antivirals, when indicated) and helps us figure out transmission patterns and risk factors to prevent more infections. The more we understand this virus, the better we can control it, all the while trying to walk that ever-changing fine line between protection and practicality.