test

Here are a couple of quick H5N1 flu updates about which I’ve been wanting to write. Neither is very surprising, but they highlight (again) some issues of concern.

(Presumed) fatal H5N1 influenza infection in a cat

An outdoor cat in Washington State tested positive for H5N1 influenza. Spillover of H5N1 infection into cats isn’t new. It occurs sporadically, most often in cats with outdoor access or fed contaminated raw poultry-based diets from potentially infected birds. In this case they specifically mentioned that the exposure was thought to be from contact with wild birds and not contaminated food.

Although the article doesn’t specify, presumably the cat died and was tested thereafter. That outcome seems to be the norm for cats with H5N1 flu. It’s possible that there are milder cases that go undiagnosed because a dead cat is more likely to be tested than a slightly ill cat that recovers on its own after a few days. Regardless, whether it’s almost always, most often or regularly fatal, infection in cats is bad. Trying to limit cats’ outdoor access (especially when flu is active in birds in the area) and not feeding cats raw poultry-based diets are two key preventive measures.

H5N1 influenza in cattle in the Netherlands

Back in December, I wrote about a litter of kittens from a dairy farm in the Netherlands that likely died from H5N1 influenza. At the time, none of the other animals on the farm (including the dairy goats and other cats) tested positive for flu. More recently, it was reported that one cow on the farm tested positive for H5N1 flu antibodies, which is consistent with the cow having been previously infected. The latest update is that five cows on the farm have now tested positive for H5N1 flu antibodies. That’s noteworthy for a few reasons:

  • The infection went relatively unnoticed in cattle. Whether they were all completely healthy or only mildly ill such that it wasn’t flagged isn’t known. This demonstrates that cattle can have relatively silent infections that may be difficult to detect clinically. It also shows (again) that cats may be great sentinels on farms, i.e. unexplained cat deaths should trigger an investigation to see if flu may be present.
  • Five positive cattle also gives us more confidence in the diagnosis. A false-positive test in five cows is much less likely than a false-positive in one. It also fits with some degree of cow-to-cow transmission. It’s unlikely (but not impossible) that there were five separate bird-to-cow transmissions, but much more likely that the virus went bird-to-cow-to-cow, as has been seen on US dairy farms. As I’ve mentioned many times before, any mammal-to-mammal transmission of this virus is a concern in terms of increasing the virus’ potential to jump to and spread among humans.

I periodically get calls from concerned veterinarians (none recently from Ontario, thankfully) along the lines of “We did surgery on a dog with a liver mass that was diagnosed as Echinococcus multilocularis (EM). What is the risk to the staff and what do we do to decontaminate the clinic?”

Depending on how well I know the veterinarian (and my mood), I sometimes start my reply with “Unless your staff are secretly canids AND snacked on the dog’s liver, everyone will be fine.”

The reason for this is based on one key aspect of the biology of EM: infection in dogs can occur from two different life stages of the parasite. (For more basic background on EM (aka the fox tapeworm), check out our Worms & Germs Echinococcus infosheet, or the Ontario Animal Health Network EM infosheet).

Dogs are primarily definitive hosts for EM:

  • As definitive hosts, dogs harbour adult worms in their intestinal tract, which are effectively harmless to the dog (they are very tiny!), but result in parasite eggs being passed in the dog’s feces, which are infectious to other animals, including people.
  • This form can also occur in other canids (e.g. foxes, coyotes) and rarely in cats.
  • Dogs get this form of infection by eating an intermediate host (usually a rodent) that was harbouring immature tapeworms in the form of cysts its body (known as alveolar echinococcosis (AE).

Sometimes dogs develop alveolar echinococcosis, which is the form of EM infection typically seen in intermediate hosts:

  • Intermediate hosts develop AE after ingesting E. multilocularis eggs from feces of a definitive host.
  • The immature form of the parasite migrates through the body and ultimately forms large budding cysts, most often found in the liver, that can spread like a malignant tumor. The disease can be fatal.
  • AE can occur in a wide range of species, including people.

It is relatively unique that dogs are a definitive host but they can also develop the intermediate host form of the infection, and that’s where the confusion typically lies. Only one form poses a risk to people.

  • A dog with intestinal infection sheds tapeworm eggs in feces which are infectious to people and other animals that ingest the eggs.
  • The AE form is only infectious to other definitive hosts (generally dogs and other canids) and only if they eat parasitic tissue from the dog. A person could snack on a raw sample of that dog’s liver and they’d be fine (at least in terms of the parasite, but still… yuck).

We can’t say dogs with AE are completely zero risk, because if the dog has been exposed to EM eggs, then it’s also possible the dog could have ingested an infected rodent with the intermediate form of EM in the same area, and may therefore have a concurrent intestinal infection. We really can’t say zero risk for shedding of EM eggs in any dog that lives in an endemic area (including southern Ontario) and spends time outdoors (because dogs will be dogs…). The fact that the dog has AE doesn’t influence that risk or make it higher risk than any other local dog.

We often test or empirically treat dogs with AE with praziquantel (an anti-tapeworm drug that is effective against the adult worms only) out of an abundance of caution, but other dogs from the same area (and with the same lifestyle risk factors, like catching rodents) are likely at just as much risk.

So… if you’re in contact with a dog with AE, relax. If you’re in contact with a dog with intestinal EM infection that may be shedding parasite eggs, then don’t eat poop and you’ll be fine. In Ontario, EM (in any form) in a definitive host (i.e. cat or dog) is immediately reportable to the local public health unit, and they will assess the risk to anyone who has been in contact with the dog (or its feces). Understanding and differentiating the different types of infections, how they occur and the different risks they pose is a key aspect of EM control and communication.

Life cycle diagram below from US CDC DPDx website.

test

A few months ago, there was a cluster of kittens from a dairy goat farm in the Netherlands that likely all died from H5N1 influenza (only one kitten was tested). That wasn’t too surprising since the H5N1 virus is circulating in wild birds internationally, and we know that cats are highly susceptible to the virus, and can develop very severe (and not uncommonly fatal) disease. Being from a dairy goat farm, there was concern the kittens could have been infected via milk from the goats, because that’s what’s been seen / suspected in cats on dairy farms with infected cows in the US. When the H5N1 virus spilled over into daily cattle in the US in 2024, it spread rapidly and widely amongst dairy farms, often infecting (and killing) the farm cats. Initially it was reported that the goats and several other cats on the Dutch farm were also tested at the time and were negative for H5N1 flu. It was therefore assumed that the kittens had been exposed to the virus via wild birds, possibly indirectly through the mother cat.

More information about this case has recently been released. It turns out the farm was a dairy cattle farm (or at least had cattle as well as goats). When they tested the cattle – which were all clinically healthy – they found one cow that had antibodies against H5N1 influenza. If the test isn’t a false positive (which is always a consideration with rare outcomes and antibody tests), that would mean at least one cow was infected with the virus at some point, but wasn’t sick enough for it to be noticed, and then eliminated the infection without spreading it to the other cows, and leaving antibodies in the bloodstream as evidence.

It’s a good news/bad news situation.

The H5N1 influenza outbreak in dairy cattle in the US has been a mess, both in terms of the impact and the approach to control. We’ve been watching to see if the virus would spread or cross over into cattle again in another country. Sporadic wild bird-to-mammal transmission of the virus in wildlife especially is not uncommon, but the infections don’t lead to further mammal-to-mammal spread. Hopefully that’s what happened on this farm. However, every time H5N1 flu jumps into a mammal, it creates more opportunity for the virus to adapt to spread more easily in mammals, or the strain that jumps could already be adapted to this kind of spread. The more domestic animals get infected, the greater the risk of human exposure, and the more risk of further adaptation or recombination with human flu viruses to make another strain that could be worse for us.

The good news is that this seems to have been a one-off situation on this farm. The virus spread to at least one cow but didn’t cause overt disease, didn’t seem to spread further in cattle, and the situation seems to have resolved itself with no damage beyond the initial cluster of infected kittens.

But this is yet another reminder of why we need to pay attention to H5N1 flu and try to control it as best we can. This situation didn’t cause a problem, but a nationwide outbreak in dairy cattle will likely start with a single spillover on a single farm, so we need to remain vigilant so that we can hopefully stop it before it spreads further from there.

test

Whether it’s dealing with an individual patient or developing international antimicrobial use guidelines, one of the bigger challenges we face in regard to antimicrobial treatments for dogs and cats is determining how long different infections need to be treated.

Unfortunately the treatment durations most commonly used are not evidence-based. We have very little data to go on, but well-established dogma that’s often largely based on something someone said or wrote (without evidence) possibly decades years ago. (This phenomenon is also not unique to veterinary medicine!)

It’s not that the traditional (old) approaches are inherently wrong, but we have to realize they’re not necessarily right, and we have to be open to change. Most antimicrobial treatment durations currently used in dogs and cats (and sometimes other species) are probably way too long; when we compare them to what’s done in human medicine for comparable conditions (for which they have a lot more large-scale studies), some of the differences are striking. But people are often (understandably) reluctant to change, even when there’s little or no evidence to support current practices, and we have circumstantial evidence (and basic principles) that suggest we’re treating patients for too long. We’re afraid of the risks of not treating long enough, but treating for too long also has risks, especially when it comes to antimicrobial resistance.

As the clinical antimicrobial treatment guidelines field advances, we’re underpinning new guidelines with evidence syntheses, particularly systematic reviews, to help change people’s minds and ultimately (hopefully) prescribing practices.

A new systematic review entitled “Shorter versus longer durations of antibiotic treatment for pneumonia in dogs and cats: a systematic review and meta-analysis” (Emdin et al. 2025) should help us revise our treatment guidance for pneumonia in dogs and cats. Veterinarians sometimes make recommendations for very long treatment courses for pneumonia in pets, which is in stark contrast to what’s done in humans and other species such as cattle.

Unsurprisingly, the systematic review only found a small number of studies (3) for dogs and none for cats. Two of the canine studies were randomized controlled trials (yay!) and one was an observational study (which provides lower level evidence and is more prone to bias, but still important to consider since we have such limited data in veterinary medicine). Meta-analysis showed no significant difference in outcomes (both clinical cure and recurrence) between animals treated with shorter (10-14 days) versus longer (21-28 days) courses of antimicrobials (see tables below or the full review for more details).

BUT there are lots of provisos. There were only 3 studies, and they were fairly small (only 74 dogs in total), so the assessment of the certainty of evidence was deemed very low (details in table below).

Despite the limitations in the data, this study suggests that we can use shorter treatment durations for pneumonia in dogs. It also highlights yet again that we need more data and better studies. We also need to study even shorter treatment courses to see if they’re effective. In humans with community-associated pneumonia (i.e. not acquired in hospital), there are numerous controlled trials showing 3-5 days of antimicrobials is just as effective as longer courses (check out Dr. Brad Spellberg’s “Shorter is better” website for more details). In cattle, we also we use much shorter treatment courses, typically only a few days, or a single dose of a long-acting drug that provides coverage for 3-8 days.

While this review suggests that we don’t need to treat pneumonia is dogs for longer than 10-14 days, but doesn’t tell us if even shorter courses would be just as effective. I suspect that if we had a study looking at 5 versus 14 days of treatment, we’d see they are equivalent. My current standard recommendation for treatment duration in these animals is 5 days; although this is admittedly based on no direct evidence in dogs and cats, we have enough indirect evidence that I’m comfortable with it as long as the patient responds clinically over those first 5 days.

Fortunately, the fields of evidence synthesis and guideline development are advancing in veterinary medicine. Unfortunately, we have many, many areas in which data are still limited or non-existent. That means we have lots of low-hanging fruit for future studies; more data will help us make more robust and reliable guidelines.

test

We call rabies “almost invariably fatal” in people. Rabies kills an estimated 50,000 people a year globally, mostly in Africa and Asia. Even with very intensive care, the prognosis is grave. Only a very small number of people have survived rabies: there are approximately 34  documented cases of survival, but an even smaller number of people have survived without serious longterm neurological deficits.

In 2004, a treatment regimen coined the “Milwaukee protocol” was used to successfully treat rabies in an unvaccinated 15-year-old girl (Jeanna Giese), after she was infected through direct contact with a bat. It involved giving anesthetic drugs (ketamine and midazolam) to put her in a coma and slow down her brain function, and two antivirals (ribavirin and amantadine), along with very intensive care to support her vital body functions. The idea was to try to protect her brain and try to reduce the rabies virus levels long enough to give her immune system time to fight off the infection. Remarkably, she survived. Even more remarkably, while her recovery was prolong, she ultimately seemed to have limited longterm neurological problems and 20 years later is a mother of three.

The success of the protocol was published (Willoughby et al. 2005) and it became the foundation for future treatment attempts. Unfortunately, while it attracted a lot of attention and optimism, its success has not been reproduced. That has lead people to question whether Jeanna survived because of the treatment, or simply because of a very fortunate confluence of exceptional factors (that we don’t know) and intensive supportive care, which resulted in her body being able to fight off the virus with few lingering effects, despite not having been vaccinated.

A recent commentary entitled Demise of the Milwaukee Protocol for Rabies (Jackson et al. 2025) highlights these issues, including that there have been no subsequent proven successes (anecdotes, but no hard evidence) and at least 64 documented failures, and that critical care is likely the most important component, as opposed to the specific drug cocktail that was used in Jeanna’s case. The author calls for abandonment of this protocol (which has been echoed by others too) and consideration of new treatment approaches.

The letters to the editor section got a little testy, with a response from Dr. Willoughby (the doctor who oversaw Jeanna’s case and published the Milwaukee Protocol), including a  disappointing comment akin to “well, how many rabies survivors have you had?”   Willoughby claims there are more survivors, but provides no details or links to any peer-reviewed data. Now not all useful data are published, so we shouldn’t dismiss the response based solely on that, but for such a high profile disease and for a protocol that’s been challenged over the past few years, if there were solid data it would be strange for none of it to be published. There might be more evidence of patients surviving, or there might just be weak anecdotes, incomplete stories, questionable data and survival of patients that didn’t actually have rabies – all of which are mentioned in Dr. Jackson’s response to Dr. Willoughby).

Dr. Willoughby states “We do not change a successful protocol even if mechanistically mysterious.” But while we shouldn’t dismiss things that might work out of hand, but we need to objectively assess them, and make sure we are not blinded by single successes, personal biases or hope.

  • Sometimes we don’t understand things that work. We don’t want to make clinical decisions solely on mechanistic aspects and proxies that don’t necessarily apply in the patient.
  • At the same time, we shouldn’t perpetually use treatments that are lacking solid proof of efficacy. We have lots of unsubstantiated dogmas in medicine (both human and veterinary) that persist for decades because they are not critically evaluated, despite no evidence to support them or even evidence against them.

In the case of rabies, one might think “Even if it doesn’t help, it can’t hurt, and we should try something.” That’s understandable, but if people think we have a possibly effective treatment, there’s less impetus to develop and evaluate other treatments that might be more effective. Clinging to a futile treatment can be harmful.

Is the Milwaukee Protocol futile?  Medicine seems to be leaving this protocol behind, but we can still learn from it, and some of the concepts remain potentially useful. Efforts to control the neurological impacts and providing intensive supportive care to keep the patient stabilized while their body gradually fights the infection are probably still key. However, the reliance on the specific drug cocktail used in the Milwaukee protocol, which has not necessarily worked apart from that first case, might be stifling more research.

Twenty-two years from this highly published success, rabies remains almost invariable fatal, and successful treatment is beyond the grasp of the vast majority of people who get infected, as most are in resource limited areas where the degree of required intensive care is not available. The search for an effective treatment continues, and hopefully we’ll find one someday that will be accessible across the world, and not just to those who can access (and afford) highly specialized care.

More important is rabies prevention. That’s still our first and most important defence against this deadly disease, including vaccination of domestic animals, vaccination and sterilization campaigns in areas where canine rabies is endemic, education to avoid bites and how to respond to bites, and improved access to and uptake of rabies post-exposure prophylaxis. Like most problems, prevention is better than treatment, but we still need an effective treatment as rabies isn’t going away.

test

Antimicrobials are often used at the time of surgery, but it’s widely accepted that there is tremendous overuse of antimicrobials in this context in both human and veterinary medicine. Antimicrobial prophylaxis is indicated in some surgical patients to reduce the risk of surgical site infection, but in a large percentage of cases use of antimicrobials is actually unnecessary and is based more on habit or fear (i.e. more to make the surgeon or pet owner feel better, versus actually helping the patient).

Clinical guidelines are an advancement in care, and the field of antimicrobial guideline development has progressed significantly in recent years. We’ve moved from primarily expert-opinion-based guidelines to evidence-based, structured guideline development, which should yield stronger, less biased and more defensible guidance, but guidelines are never perfect, since there are typically still lots of evidence gaps. Also, guidelines are meant to cover the majority of situations, not every possible case, so there are always exceptions to the “rules.” Nonetheless, good guidelines help support good clinical decisions.

The European Network Optimization of Veterinary Antimicrobial Therapy (ENOVAT) has just released their new 2025 guidelines for surgical antimicrobial prophylaxis in dogs and cats. These guidelines are the culmination of several years of work, and are underpinned by a thorough scoping review of antimicrobial prophylaxis in companion animal surgery (Sorensen et al. 2024) and a systematic review on the same topic (coming soon).

The guidelines have a heavy emphasis on when NOT to use antimicrobials, since that’s what the evidence supports, but they also highlight situations where antimicrobials are recommended, and they provide details about optimal approaches for when they’re needed.

The guidelines used a GRADE-based approach, which culminates in a strong or conditional recommendation for or against each intervention (or a conclusion that we can’t make a recommendation either way).

For strong recommendations, the guidelines say “we recommend…”

  • Strong recommendations for an intervention (i.e. antimicrobial prophylaxis) are made based on moderate to high certainty evidence of the effect of the intervention, plus supporting value amongst various other domains (e.g. importance of the problem, benefits, harms, cost:benefit, equity, acceptability).
  • Strong recommendations against an intervention can be made based on similarly moderate-high certainty evidence or where there is lower certainty evidence about the effect but moderate/high certainty evidence about potential harms. The default is not to use the intervention, so a strong recommendation against can be made without solid data showing it doesn’t work.

For conditional recommendation, the guidelines say “we suggest…”

  • Conditional recommendations are made when the recommendation is based on low certainty evidence or when there is a lot of uncertainty or variability about acceptability, applicability, equity or other factors that indicate it might not be an ideal or preferred approach for most. (For example, we’re not going to make a strong recommendation for a treatment that’s so expensive only a small subset of the population could ever use it.)
  • A conditional recommendation doesn’t mean that the treatment is less effective than one with a strong recommendation, it simply means the confidence and certainty behind the recommendation is lower, and/or that its value is more situational, being a preferred choice in some situations but not others.

Below is a quick synopsis of the recommendations (in tiny print). Check out the full text article (which is open access) for more details on each recommendation and the evidence behind them all.

test

We’re in the midst of a pretty bad human flu season. That’s a problem by itself, as it also means severe flu cases and hospitalization rates are high, and likely to increase. High flu activity in people also amplifies concerns regarding H5N1 avian influenza, because it creates more opportunities for an infection with both a human flu virus and the H5N1 avian flu virus to occur in the same person (or animal) at the same time. That could lead to recombination of the two viruses creation of a new flu strain that is more severe and transmits effectively person-to-person.

H5N1 avian flu has spilled over into a large number of mammals. Among domestic animals, cats and cattle have attracted the most attention, but there is also some risk to dogs. The number of infections in dogs has been very low (especially considering how often dogs are likely exposed to infected wild birds, other wildlife, cats or livestock), but in a couple of cases infection has been shown to be quite serious. Case in point was a recent fatal H5N1 infection in a dog in Alberta.

In November 2025, a 10 year old golden retriever / poodle cross dog (better known as a goldendoodle) was infected with H5N1 influenza after exposure to a snow goose, and subsequently died. The dog had a compromised immune system (it was being treated for an immune-mediated disease), and that could have contributed to severity of the disease caused by the flu virus. It’s similar to the case that occurred in an Ontario dog in 2023, that died from H5N1 flu after being exposed by chewing on a Canada Goose that died from the same infection.

This case is more of a reminder of the risk that’s already been present for a while, versus anything new, but it’s part of the reason we put a lot of effort into H5N1 influenza control. Currently circulating H5N1 flu strains are poorly adapted to infect and spread between people (and dogs, and most other mammals with a few exceptions), but every time the virus spills over into a mammal, it creates more opportunity for the virus to change and adapt to infect more mammals, which is definitely bad.

test

That headline might get some people worked up, but hopefully they’ll read the whole post before firing off an angry email.

Antimicrobial resistance (AMR) is a huge problem in humans and animals, which means we need to improve or optimize how we use antimicrobials, but this is not synonymous with reducing use. Most of the time, improving and optimizing antimicrobial use (AMU) does focus on reduction, because we greatly overuse these drugs in general. Most of the national and international discussion around AMU also focuses on reduction, typically based on one quite crude measure: mass (kgs, tonnes) of drug used in animals over time (e.g. annually). Overall mass of drug used is usually one of the easier numbers to get (though even that can take considerable work depending on the system), and sometimes its useful, but it’s not always the best. Other times it’s simply not helpful, and other times it’s downright misleading.

I wrote a more detailed commentary on this same topic that was recently published in the Canadian Veterinary Journal, but here is a quick summary of some of the reasons why we need move beyond looking at simply reducing the mass of drugs used when it comes to improving antimicrobial stewardship.

All antimicrobials are not created alike

We have a lot more concerns about resistance to some drugs (higher tier) compared to others (lower tier). That’s why antimicrobials used in animals are categorized based on their potential impact on AMR in humans. Simply measuring the mass of all the drugs used together doesn’t consider which drugs are being used. It’s also important to account for differences in dosing. Newer, broader spectrum, higher tier drugs tend to be more potent, so they are used at lower doses (see table below), but the consequences of resistance to these drugs is much higher.

Based on this, we could significantly drop the mass of antimicrobials used in animals by using more higher tier drugs in place of lower tier drugs – but that is exactly what we don’t want to do! If a country is pressured to reduce AMU, the risk is they may simply substitute higher tier drugs for lower tier drugs, and then say “look how much we’ve reduced our antibiotic use, good for us!” while ignoring the disaster in the making.

Related to this, an increase in the mass of antimicrobial use might be good in some situations. Let’s use dogs with urinary tract infections as an example. In the table below, the “pre-intervention” column is based on 1000 dogs treated as per a study of AMU practices in dogs with urinary tract disease in the US and Canada. If we implement an intervention to move practices to 100% compliance with ISCAID guidelines for treating urinary tract disease in dogs, it would increase the use of three lower tier drugs, which is good. But, it would also increase the overall mass of drugs used by quite a bit, based on how they are dosed. Looking at the effect on mass of drugs used alone makes this look like a harmful intervention, when in fact the change would be good from a stewardship standpoint, since it would mean using the lower tier, recommended drugs.

Increased animal production

The world is growing. Populations in some countries are expanding greatly, and overall wealth is increasing, leading to more need for food, and more demand for meat. We can debate the ecological aspects of meat production, but ultimately we can’t expect countries to stop increasing their own food production to satisfy a numerical target for antimicrobial use. If a country expands its meat production by 100% but only increases their AMU by 25%, it means they are using less drug per animal or per kilogram of meat. That’s a step in the right direction and a big win in my mind, as it’s often accomplished through improving overall animal health, which provides many other benefits beyond improved AMU. If a country is expanding meat production, they’re not going to sign onto any agreement that limits the overall mass of antimicrobials used, but instead we should be working together to optimize animal health in order to optimize antimicrobial use, so we maximize the benefits and minimize the risks.

Increased access to antimicriobial drugs

“Use as little as possible but use enough”. That’s my mantra when it comes to antimicrobials.

We can use too little. Animals get sick, and sometimes they need antibiotics (just like people). Internationally, a lot of animals that would be helped by an antibiotic don’t get one because the antibiotic is not available or not affordable, or there’s limited or no understanding of how/when to use it. This is most common in low income countries, but access issues in different countries vary. Improvements in drug availability, less poverty, increased literacy, improved transport infrastructue (e.g.roads), better access to veterinary care and other factors can reduce underuse of antibiotics. That can lead to a short term increase in antibiotic use, but it’s better for animal health and welfare, and hopefully is ultimately combined with improvements in animal management, preventive medicine, vaccination and other basic care.

All that said, we still need to reduce antimicrobial use

We massively overuse antibiotics (in animals and in people). However, AMR is a complex problem, and complex problems can’t be addressed with simplistic, sound-bite approaches, like picking a random number and an arbitrary date and saying “let’s reduce antibiotic use in livestock by X% by the year 20XX.” We need more nuanced discussions to maximize animal health, use antibiotics when needed, choose the right drug and duration in each case, and stop using them when they are not needed. Some of that will reduce AMU, but some might increase AMU (at the mass level or even overall), but ultimately improve overall health of animals (and people) and reduce the impacts of AMR.

We also need to think about what we really want to measure. As you can tell, I’m not a fan of measuring overall mass of drugs used. It’ better than not measuring use at all, and it can be useful in some situations, but it’s mainly used because it’s a relatively easy number to calculate. We’re better off focusing on metrics for appropriate use of antimicrobials, such as:

  • The percentage of animals of a certain age or health status that are treated with an antibiotic therapeutically or prophylactically
  • The percentage of treatments that are consistent with available AMU guidelines
  • The percentage of treatments that use lower tier drugs
  • The percentage of animals that are treated based on the advice of a veterinarian

These metrics take a lot more time to develop but are more understandable, more actionable and ultimately lead to better surveillance-to-action, and impact.

test

Change is tough. Repeated changes are even tougher, especially when it takes a lot of time and effort for each one to be understood and implemented. But change is also good – and important – when it improves how we do things.

In 2024, the Clinical and Laboratory Standards Institute (CLSI) changed some important breakpoints for antimicrobial susceptibility testing in dogs. Breakpoints are what labs use to to report whether a bacterial isolate is susceptible or resistant to different antimicrobials. As we learn more about the bugs and the drugs and how they interact, sometimes those breakpoints need to be adjusted, so that the veterinarian receiving the report is getting the most accurate information possible to make better treatment decisions for the animal. That’s good. But it also involves change, which can be slow. It’s taken close to two years for some labs to fully implement the 2024 changes.

Now we have data suggesting more things need to change. CLSI hasn’t updated any mrore breakpoints (yet), but published data indicate that we should probably be rethinking how we’re interpreting certain antimicrobial susceptibility test results for cats.

A recently published study (Papich et al. 2025) came to the conclusion that that breakpoints for fluorquinolone antibiotics for some bacterial isolates from cats should be lowered. That means some bacteria that would currently be reported as susceptible should actually be considered resistant. It’s very similar to the changes that were made in 2024 for isolates from dogs; I suspected at the time a similar change would eventually be needed for feline isolates, but it hadn’t yet been studied enough.

The researchers examined pharmacokinetic data, pharmacokinetic-pharmacodynamic (PK/PD) analysis and susceptibility data from a large number of bacterial isolates to determine the appropriate breakpoints for enrofloxacin and marbofloxacin when it comes to Enterobacterales (E. coli and related bacteria), Pseudomonas aeruginosa, Staphylococcus spp., Pasteurella multocida and Streptococcus. Skipping to the punchline, what they found indicated that we should be using lower breakpoints for these bug-drug combinations in cats. The current breakpoints would classify certain isolates as susceptible to these drugs, when in fact the likelihood of achieving adequate inhibitory drug levels in the target tissues using standard dosing was very low.

Here is a summary of the old (technically still current) and suggested new breakpoints for these bug-drug combinations in cats:

They also suggested a “susceptible dose dependent” (SDD) breakpoint, as was done for dogs. This means that the bacterium can be considered susceptible when a higher dose of the drug is used in the patient. For marbofloxacin, isolates that meet the SDD breakpoint are only considered susceptible when the cat is dosed at 5.5 mg/kg (the high end of the label dose for this drug). There are no SDD breakpoints for enrofloxacin because we don’t want to use higher doses in cats due to the risk of causing blindness (retinopathy). Personally, I have no use for enrofloxacin in cats at all because of this risk. but if it’s going to be used, we don’t want to go beyond 5 mg/kg regardless, so there’s no SDD breakpoint.

What do we do now?

  • We have data suggesting that we need new breakpoints, but the CLSI standards haven’t yest been changed.
  • Personally, I’ll start using the new breakpoints right away. I trust the research group and the data, and the new breakpoints are consistent with the change that was made for canine isolates.
  • When we get MIC data directly on our lab reports, it’s easy to apply the breakpoints ourselves. The challenge is when labs only test a narrow range of drug concentrations or don’t report MICs (i.e. the report doesn’t provide the number, just susceptible intermediate or resistant (S-I-R)). Then we’re a bit stuck, and my confidence in using enrofloxacin or marbofloxacin would decrease.

Pradofloxacin, the newest licensed fluoroquinolone in cats (and dogs in some countries), was not included in this study. The breakpoints for this drug were not changed for dog isolates in 2024 either. Whether that’s because the newer breakpoints for pradofloxacin are fine, or there isn’t enough data to re-assess them, or it was just lower priority to study isn’t clear, but it raises questions. If I see an isolate that is resistant to marbofloxacin and enrofloxacin based on the new breakpoints in dogs (or the suggested new breakpoints in cats), and it’s classified as “susceptible” to pradofloxacin but right at the breakpoint, I’d be wary of using it. It might work perfectly fine, but it gives me a bit of pause, and I’d be inclined to look at other options.

I’ve been meaning to write about this for a few days, but a case from today prompted me to finally do it. It was a cat with an E. coli infection that was reported as susceptible to fluoroquinolones, but looking at the MICs and this paper, I’m not confident that it actually is. So, I recommended a different drug. I actually would have recommended the different drug anyway, since it was a lower tier option that should nonetheless be effective, but this paper changed my assessment; had my options been more limited, I would have searched for a non-fluoroquinolone option.

This paper and the new suggested breakpoints add more complication to an already complicated area, but while it can be a hassle, and change is still tough, it’s important progress. It will help veterinarians provide more effective patient care and improve antimicrobial stewardship, by avoiding using drugs that aren’t likely to be effective against specific infections in cats.

test

I’m on my way back from Copenhagen where we had a very productive meeting to update the ISCAID pyelonephritis antimicrobial treatment guidelines for dogs and cats. As the process for developing guidelines like these has matured, it’s no longer about simply getting some very smart people in a room and agreeing on recommendations; it’s now a much more structured, evidence-based process. As part of that, we think about more than just “would this drug work?” We also think about factors like adverse effects, cost, acceptability, feasibility, equity and others. One of the newer consideration is now “planetary health,” which is applicable to a lot of things, including antimicrobial use guidelines.

Antimicrobial production, distribution and use have carbon footprints and require other resources that impact more than the individual who gets treated with the drug. (And yes, I fully recognize the irony of talking about carbon footprints while flying across the Atlantic in a plane, but sometimes in-person meetings are important too.) While we’re not going to dramatically alter our guidelines based on a drug’s carbon footprint, it’s something we need to at least think about for awareness. A side benefit of good antimicrobial stewardship resulting in less antimicrobial use is smaller footprints of this sort. But, are those footprints really relevant? It’s always hard to figure out what the contribution of something like a drug (or a flight) is to the big picture, and individual events have near negligible impacts. But, when we do something over and over and over again, the cumulative impact starts to become more relevant.

What do we know about the ecological impacts of antimicrobials? I’m far from an expert in this, but it’s interesting, so I’ll just toss out a few points – food for thought.

That’s a random collection of studies on the topic, and ultimately we don’t really know the full downstream effects of antimicrobial use, but it’s fair to say that these drugs have a big carbon footprint, and we can reduce it through antimicrobial stewardship: using fewer antibiotics, using them better when necessary, and, most importantly, optimizing health so we don’t have sick people or animals to treat in the first place.