test

Two separate stories regarding Salmonella contamination of unrelated pet products have been in the news lately.

The most recent one was a recall of Salmonella-contaminated pet supplements from Fi, a New York based company (see the link for specific products and lots affected). Details are pretty sparse, but the recall was apparently triggered because of contamination of an ingredient supplied to the company. It’s not clear whether the contamination of the ingredient was detected by the supplier during routine testing, by Fi during routine testing, or based on an investigation linked to sick pets. The FDA notice doesn’t mention anything about sick animals (or their human contacts), so hopefully this recall was made out of abundance of caution prior to any known cases of illness (while noting that “no known cases” doesn’t mean “no cases”, as illness from contaminated food, treats or supplements is vastly underidentified and underreported).

While contamination is concerning, the response seems very appropriate: A problem was identified, the FDA was notified, and a recall was initiated. That’s how it should be.

…But at the other end of the spectrum is Darwin’s Pet Food.

I’ve written about Darwin’s before as they’ve had multiple contamination problems with their raw pet food in the past, including E. coli in contaminated pet food resulting in severe disease in a child, follow a few months later by more issues with contamination of their products with Listeria and Salmonella. A common theme with both stories (beyond the biohazardous pet food) was the company ignoring FDA requests to recall the implicated diets. It boggles my mind that a company can just ignore an FDA recall request, both based on the legal aspects and basic ethics. But, they did.

More recently, the US Justice Department filed a complaint to “to permanently enjoin a Washington state pet food manufacturer [Darwin’s] from manufacturing and selling adulterated pet food.”

It shouldn’t take a Justice Department lawsuit to make a company stop selling contaminated food but that seems to be the case with Darwin’s. In the FDA notice, Timothy Schell, Director of FDA’s Center for Veterinary Medicine stated “Despite repeated FDA warnings, this company continues to manufacture and distribute products contaminated with harmful bacteria. When a company does not take responsibility for product safety, FDA will intervene to protect public health.”

In some ways, I’m surprised they acted, as contamination of raw pet food is a niche issue that has largely avoided regulatory attention in the past. It’s a good sign, though. Our archives have lots of information on the broader issues around whether or not feeding raw diets to pet is a good idea, so I won’t get into those again here. Suffice to say, if raw diets are fed to pets, people need to realize there’s an increased risk to both pets and people, and they need to take appropriate precautions. Companies need to do that too. For example, some raw pet food companies use high pressure pasteurization to reduce (not necessarily eliminate) bacterial contamination in their products, which is a very useful measure.

Good companies have good facilities and food safety practices. Others deflect responsibility and ignore warnings and reports of sick animals or people. The problem is, consumers often can’t tell which companies are doing (or not doing) what. Government action is important in this kind of scenario in particular; naming-and-shaming can be a big part of it since, ultimately, if consumers stop buying dodgy products, those companies will either go out of business or have to improve.

Here’s some other info from the Justice Department notice:

The complaint, which was filed in the U.S. District Court for the Western District of Washington, alleges FDA found the presence of pathogenic bacteria, including Salmonella, Listeria monocytogenes (L. mono), and Shiga toxin-producing E. coli (STEC) in the company’s finished pet food products in samples collected from 2017 through 2025. In 2024, FDA investigators also found Salmonella in the company’s Tukwila facility. Infections from L. mono, Salmonella, or STECcan cause symptoms such as diarrhea and vomiting in healthy adults. For vulnerable consumers — including pregnant women, the elderly, and the immune-compromised — L. mono, Salmonella, and STEC can each cause more serious effects including death. L. mono can cause stillbirths and miscarriages.  

The complaint further alleges Arrow Reliance’s products have repeatedly infected consumers over several years. After multiple Arrow Reliance customers complained of humans or pets suffering from health problems, FDA investigators detected the presence of pathogenic bacteria in unopened samples of the company’s product. In 2024, a four-year-old child fell ill with a STEC infection and developed Hemolytic Uremic Syndrome after the family dog was fed Arrow Reliance’s products. Third party laboratory testing of the family’s unopened pet food from Arrow Reliance confirmed the presence of both Salmonella and STEC bacteria. 

Here we have the juxtaposition of one company that seems to have initiated a recall of its products because an ingredient was contaminated, without evidence of disease but acting appropriately to reduce the risk (yay!), compared to another company that has chronically and repeatedly ignored FDA recall requests, has clear evidence of long term quality control and contamination issues, and shows no interest in the health of consumers or their pets (boo!).

I hope that people who feed Darwin’s diets will pay attention to what this government action is saying.

test

Antibiotics get used a lot in dogs and cats (and other species) that undergo surgery. A lot of the time, the antibiotics are unnecessary. I suspect in many cases veterinarians know it’s unnecessary, but they still use them out of habit, fear of complications, fear of complaints from owners, lack of consideration of the potential downsides, and because we are programmed to want to “do something” to avoid adverse outcomes, even when not doing anything may be the best approach.

A big challenge with developing reliable veterinary antimicrobial prophylaxis guidelines is the lack of good studies. There are lots of small observational studies in animals for various types of surgery, which are useful but only provide quite low-certainty evidence. We’d love to have randomized controlled trials for each type of surgery in each species, but such trials are expensive and complex. Another major challenge for any surgical site infection study is enrolling enough cases (i.e. sample size). Since surgical site infections are uncommon after most procedures (which is a good thing!), studies usually need very large sample sizes (e.g. hundreds to thousands of animals) to detect differences between treatment groups, or to confidently say there’s no difference (i.e. non-inferiority trial). I frequently have discussions with people who want to do studies looking at use of antimicrobials and surgical infections. Once we go over the numbers, they tend to quickly get turned off because sample size calculations show the required size would be way too big for what they can manage (or afford) to do. Unfortunately, there simply is not enough funding in this area, so we’re not likely to get these kinds of large trials anytime soon. That doesn’t mean we’re stuck with no evidence at all on which to base our guidelines, it just means we have to rely on different types of evidence that have less certainty.

That can be disheartening to feel like the kinds of studies we want to do are constantly out of reach, but it doesn’t mean that smaller studies are futile even if they can’t answer all the questions. “Don’t let perfection be the enemy of the good” as they say. We just have to have realistic expectations (and avoid over interpreting the results).

We recently published a commentary highlighting this, entitled Small sample sizes in clinical trials: a pragmatic approach to clinical research in veterinary medicine (Weese et al. J Small Anim Pract 2026). A single small study may not be able to answer our big questions, but we can put data together from multiple small studies (if they’re well designed) using meta-analyses to draw stronger / broader conclusions. The concept is that there are no underpowered studies, there are only underpowered analyses.

  • Small studies may not be amenable to much or any statistical analysis on their own, yet weak or futile analyses are often attempted, likely because the authors feel it’s expected. However, improper conclusions from underpowered analyses can range from useless to even harmful in some cases.
  • We still want small studies to be published, but they may just be data with no analysis, and that can be hard for researchers and readers alike to wrap their heads around.

I raise this concept in the context of a nice recently published little study about bacterial endocarditis in dogs that underwent balloon dilatation because of congenital pulmonary stenosis (Zeedijk & Szatmári et al. 2026). The authors evaluated dogs that underwent this procedure and that had adequate post-operative follow up, with focus on the 83 dogs that didn’t get peri-operative antimicrobials. None developed an infection. There was a smaller group of 11 dogs that did antimicrobials. None of them developed an infection either.  We could run a rather futile statistical analysis and conclude that there is no statistically significant difference, but we’d have no confidence in that analysis. The study was not adequately powered for that comparison, so it’s great they actually did not try to do it, but we still have those data for a future meta-analysis.

The 83 dogs that didn’t get antimicrobials can also provide some additional insight. With zero infections in 83 dogs, the 95% confidence interval for the true infection rate would be 0-3.6%, i.e. the true incidence of infection could be between 0 and 36 infections per 1000 dogs that underwent the procedure without antimicrobial treatment.

  • Based on this low rate of infection, the potential severity of disease if infection occurs, the ability to treat such an infection, and the potential for complications from prophylaxis, we can consider the balance between risks and benefits of antimicrobial use in these patients. That’s still challenging, but the point is it’s important to consider all these different factors and not just the infection rate.

We can take it a step further, too. Consider that not all infections are preventable, even when antimicrobials are used, so we shouldn’t base our calculations on the assumption antimicrobial use would eliminate all these infections, it would only lower the infection rate.

  • Using an infection rate of 3.6% (the upper limit of the confidence interval) and an estimate that antimicrobials would reduce the infection risk by 25% (remembering that we don’t know if they will in fact reduce it at all), the calculated absolute risk reduction would be 0.9% (i.e. 9 infections per 1000 treated dogs).

In this scenario, the number needed to treat (NNT) to prevent a single infection is 111 dogs. (NNT is an underused but very helpful concept to put rates like this into context.) At first glance, an NNT of 111 might seem quite reasonable, especially since infective endocarditis can be a severe disease. However, that’s based on a very conservative assumption that the true infection rate is at the upper end of the confidence interval, and we have yet considered other factors:

  • Some of those 111 dogs will experience adverse effects from the antimicrobial. Most will be minor side effects, but occasionally they can be significant. So we need to weight the benefit to some dogs versus the potential harm to others.
  • Also remember that we’re basing this number on a very (and likely unrealistically) high infection rate. If the infection rate is 1% and antimicrobials still reduce that risk by 25%, the absolute risk reduction would be only 0.25%, or 2.5 infections prevented per 1000 treated dogs. The NNT would be 400. Drop that infection rate to 0.5%, and the NNT is 800.

The endocarditis rate for this procedure in humans has been reported at 0.12%, which would make the NNT 3333. I think it’s pretty safe to say we’d do a lot of damage treating a few thousand dogs to prevent one case of endocarditis (and that’s not even taking into consideration the risks of antimicrobial resistance selection). Yet, antimicrobials are still very commonly being used in dogs for catheter-based cardiac interventions that are very low risk for infection (Blok et al. 2025).

Too often we don’t do the math. We don’t think about absolute risk reduction, NNT, number needed to harm, broader risks like selection for resistance, or even cost to owners. We get tunnel vision about incidence rates and p-values, and act as though they provide the answers, when in reality they rarely tell the whole story, and are sometimes even misleading.

test

These days, it’s really easy to identify new viruses. Our technology has advanced so much that you can sequence a vast array of viruses from almost any sample from an animal, person or the environment. Finding the viruses is easy, but understanding what, if any, roll they play in the grand scheme of things is a much greater challenge.

As a reviewer for scientific journals, I see lots of papers describing new viruses. Most of these are probably old viruses that we just recently identified. Often, studies say “look what we found” but don’t take the next step to figure out if the “what” is relevant, or they tend to overstate the potential relevance. I might find a new virus in a sick animal, but that doesn’t mean the virus is causing disease; it might just be part of the normal innocuous viral population that lives on and around even healthy animals.

We don’t want to dismiss new findings, but we also don’t want to over-react. Take amdoparvovirus for example, or more specifically, raccoon dog and fox amdoparvovirus (RFAV). Amdoparvovirus in part of the Parvoviridae family, along with the more well-known canine parvovirus, but is a distinct genus that includes a variety of different viruses that infect a range of mammals, particularly mustelids (e.g. ferrets, mink), skunks and raccoons. One of the most well-recognized members of this genus is the cause of Aleutian disease, a serious disease in mink.

Raccoon dog and fox amdoparvovirus was first reported in 2014 as part of an investigation of an outbreak in raccoon dogs and Arctic foxes on fur farms in China. Disease was most severe in young animals, with signs of decreased appetite, weight loss, slow growth and chronic diarrhea, with histological evidence of gastrointestinal and kidney damage.

A recent preprint (i.e. a non-peer reviewed preliminary paper) (Gajdov et al. 2026) describes an outbreak of disease in dogs attributed to RFAV. Not surprisingly, given the world we now live in, social media in some places has taken it from “here’s a potential spillover of a virus in a single kennel” to “OH NO! A NEW DOG VIRUS!!!” This preprint describes an interesting outbreak with some nice details along with some clear gaps.

  • I try to pay attention to reports of “new” issues, but always with a healthy dose of skepticism. Most of the time, a strange disease is just an unusual presentation of a normal disease. Similarly, a “new” disease event (like RFAV in dogs) is more often a rare event or something that’s happened before but not previously been recognized, versus true emergence of a new problem.

Here’s a synopsis of the outbreak and pathogen investigation described in the preprint:

  • An outbreak of disease was identified in Dobermanns in a kennel in Serbia. Affected dogs had abnormalities such as ocular and nasal discharge, conjunctivitis and weight loss, with some dogs also having “blue eye,” liver disease or, at later stages, neurological disease (which sounds pretty textbook for canine distemper). They tested for distemper, among other things, and results were negative. Samples were submitted initially from a dog and her two puppies, but it’s not clear if only those dogs were tested for distemper or whether other dogs were as well. More details about the testing (e.g. test type, numbers, timing) would be useful to help understand if we can really exclude distemper as the cause. As the old adage goes, common things occur commonly.   However, they mention repeated testing, and presumably they did a pretty good investigation before embarking on sequencing to identify any new pathogens.
  • The authors then went virus hunting, using sequencing to look for viruses present in lung, kidney and liver samples from the dog and puppies. They found “RFAV” — more specifically, they found sequences with 97% genetic similarity to the limited number of known RFAV genomes. We know relatively little about the genetic variation of this virus in nature, which is important for interpreting this finding. If the virus is genetically similar to RFAV strains circulating in wildlife, that would support direct spillover into the kennel from an infected wild animal. If it is substantially different from viruses circulating in wildlife, that suggests that it could be a distinct variant, although that still wouldn’t mean it is a dog-adapted virus. It’s way too early to say. At this point, a spillover of RFAV into dogs seems more likely than the emergence of a distinct canine virus.
  • They then developed a PCR test specifically for this virus. They got positive results from one other epidemiologically-linked dog, and an undefined number of healthy dogs from other locations tested negative. They don’t say if they tested other affected dogs, which is important context. The more sick dogs that test positive and the more healthy dogs that test negative, the more convincing it is that the virus is clinically relevant. But even with these kinds of findings, association doesn’t necessarily mean causation.

Lack of histopathological investigation (both in general and specific testing to look for evidence of canine distemper virus) is a limitation of this paper. I wonder if they’ll get asked to do that prior to peer-reviewed publication, if they still have tissues from these animals.

Another gap is this paper is the limited epidemiological data. No information was provided about potential sources of exposure or transmission patterns. I wanted more details about the kennel, numbers of affected and unaffected dogs, a timeline of how infections developed, kennel management practices and diet, as well as potential opportunities for direct or indirect contact with wildlife. There shouldn’t be farmed raccoon dogs or foxes in the area since that’s been banned in Serbia, but wild fox exposure would be worth investigating.

In the end, we’re left asking if this situation is:

  1. A new emerging disease caused by a new virus that may lead to widespread problems in dogs..
  2. An uncommon spillover event that happens periodically but just hasn’t been identified before, because such detailed investigation of kennel outbreaks is rare.
  3. An outbreak cause by a co-infection where RFAV was perhaps a contributing factor in disease but there was something else involved too
  4.  A confluence of factors that allowed a really rare scenario to develop.
  5. A distemper outbreak that was not detected.

I’d like to have convincing details to rule out E before anything else, but by next guess would be B… that this virus is “new to us” versus “truly new”.  

I don’t want to downplay the usefulness of information though. Studies like this form the foundation for further studies to help us figure out if this virus is actually a health threat in dogs and how common it is in the population (sick or healthy). Now that this is on the radar, testing can be done (through research labs) in situations where it might be relevant. We’d also need to have some testing of healthy dogs to provide context.

Raising awareness versus causing panic is always a tough balance with infectious diseases. This report tells us we have something else to consider and to investigate further, not that we have a new concerning problem that’s causing an imminent threat.

test

I get a lot of questions about what to do when a healthy dog tests positive for antibodies to Borrelia burgdorferi, the bacterium that causes Lyme disease. It happens regularly because this test is often run alongside blood tests for heartworm, whether we want it or not – it’s really just along for the ride. A small percentage of healthy dogs have antibodies against Borrelia, so when millions of dogs get tested, there will be a lot of positive tests.

The tests are quite accurate too, but there’s a problem with how the test results are framed. Borrelia serology in a healthy dog is not a “Lyme disease test.” These tests alone can’t diagnose Lyme disease anymore than a high white blood cell count by itself can diagnose pneumonia. They just tell us the dog has encountered the bacterium and mounted a normal immune response. Most of the time, Lyme disease did not (and never will) develop. But when people refer to it and use it as a “Lyme disease test,” it leads to a lot of stress, more testing and unnecessary treatment, which can be harmful. When we get a positive test result, all too often our own insecurity makes us over-react.

What should we do when a dog tests positive for Borrelia antibodies?

Data to guide us on exactly what to do are unfortunately pretty sparse. There have been two ACVIM consensus statements on the subject (2006 and 2018), but there wasn’t a lot of consensus in the end, and many recommendations didn’t have much or any supporting evidence (That doesn’t mean they are inherently bad, it just means we need to better understand the issues). We have to look at the data we have now, what we know about infectious diseases and what we know about Lyme disease in humans. Here’s my thought process about the different options for what to do with a dog that tests positive:

Double check the dog’s health status

If we find a dog that is positive for antibodies against Borrelia, my first question is “how is the dog doing?” If the dog doesn’t have any signs that might be suggestive of Lyme disease, that’s pretty much the end of the discussion. I still want to be sure to take a very careful history to be confident that there really isn’t anything relevant going on (e.g. mild shifting lameness that could suggest Lyme arthropathy), but if the dog looks healthy, acts healthy and the owner is happy, I’m happy.

Do NOT reach for antibiotics

Unfortunately, thousands of dogs are treated with antibiotics every year because they have Borrelia antibodies, not Lyme disease. There’s no evidence that antibiotics help in dogs that have no overt signs of illness. A positive test result is an indicator of historical exposure, not necessarily an indicator of active infection. Dogs can remain positive on serological tests for years. Antibiotic treatment of a random healthy seropositive dog is very unlikely to help. Doxycycline, the typical treatment of choice, is generally safe, but there is always a risk of adverse effects. Selecting for antibiotic resistance in the dog’s microbiota (which always contains numerous potential pathogens) is also a concern. The more we use antibiotics (including doxycycline), the less effective they’ll be down the road. The cost:benefit ratio is clearly tipped towards “cost” from my standpoint.

Testing for proteinuria?

There’s no evidence that testing for proteinuria is useful in seropositive dogs either. This was recommended in the older 2006 ACVIM consensus statement (despite little consensus), but it was based on no data. I suspect it was more of “we have to suggest something so let’s suggest something that’s not harmful” versus “we have reasonable evidence to suggest this is useful.”  

The main reason for the recommendation was the concern about Lyme nephritis, which causes severe kidney disease. It’s rare but its often fatal. It also seems to be a very acute disease, in that it hits quickly and very hard. It will cause high levels of protein in urine, but the odds of us finding proteinuria as the initial sign in an otherwise normal dog (i.e. that is just about to crash from Lyme nephritis but hasn’t yet) are exceptionally low. Also, Lyme nephritis is an immune mediated disease, not an infection of the kidneys, so treatment is immune suppression. We’re really unlikely to start a healthy dog on an immunosuppressant drug, even if we do find some proteinuria.

Finding proteinuria means we need to go looking for the cause, not jump to the conclusion that the dog has Lyme nephritis. Testing for proteinuria is never bad, and it’s a very non-invasive test, it’s just of limited benefit; it’s probably no more useful in a seropositive dog than in a seronegative dog.

Skip the quantitative C6 testing

This test puts a number (quantitative) to the positive antibody result. Initially, when it first came out, we were hopeful that it would be useful if higher levels were more indicative of disease and changes during treatment could be used for monitoring response. Unfortunately, none of that has panned out. There have been no data suggesting this test provides any useful additional information, so I don’t recommend it.

Watchful waiting

“Do nothing” sounds dismissive, but it’s largely what’s indicated. It’s fair to have owners be aware that there’s a chance the dog could develop Lyme disease, and that they should watch for compatible signs of disease, but that could also be said about any dog in an area where the disease occurs.

Talk about tick exposure

This is the big one for me. A positive result means there is a need for tick prevention. If the dog is currently on a preventive, it’s worth discussing compliance since the available products are generally very good at preventing ticks from attaching for long enough to transmit Borrelia. If compliance seems to be excellent, we can add other strategies like tick checks and avoiding high risk areas to help prevent exposure. Lyme disease vaccination can also be discussed as a “plan B” approach for when ticks break through preventives. Overall, the positive result is an entry point for a discussion with the owner about the need to up their game in terms of tick control and prevention. That’s valuable for Lyme disease and other tickborne disease circulating in the area too.

Is there any value in routine testing for Borrelia antibodies?

Maybe. There are two things to consider: the patient and the population.

At the patient level, I don’t see any direct value from these tests from a disease standpoint, if the dog is healthy. However, they can be useful from an educational standpoint and initiating (or revisiting) the discussion about exposure to infected ticks and tick prevention practices, including product selection, compliance and duration of administration (the period of risk is now approaching 12 months a year in many regions since ticks are active once it gets a bit above freezing).

  • If we’re going to use these tests, we need to remarket them. They should be considered “tick exposure tests,” not tests that indicate the need for treatment or other diagnostic testing in a healthy dog. A positive result means there’s definitely been tick exposure, but a negative result doesn’t rule tick exposure out, since not all ticks (or tick bites) transmit Borrelia.

These tests are most useful at the population level. I don’t really care what the result is in a an individual healthy dog. However, when we put together results from lots of healthy dogs from lots of regions, and repeat the testing over time, it can give us a very useful picture of the distribution of evolution of exposure risks. That can help us better understand, prevent and manage tickborne disease, and better understand parallel exposure risks in people. We know much more about Borrelia exposure in dogs than in people, because we test millions of dogs over time. If we use that information wisely, it can be useful. If we use it poorly, we cause harm. I suspect we do more harm than good currently, but there’s still lots of opportunity to improve.

test

Cyclospora has been in the news a lot lately because of a large, ongoing outbreak of cyclosporiasis in people in the US. Total case counts aren’t clear (and official numbers are always lower than the true number of cases because of under reporting) but it’s estimated that well over 10 000 people have been infected.

For those unfamiliar with Cyclospora, it’s a group of protozoal parasites that can cause gastrointestinal disease. This outbreak specifically involves Cyclospora cayetanensis, which is a well known cause of diarrhea in people. Infected individuals shed the parasite oocysts in their feces, but it takes 1-2 weeks in the environment for the oocysts to sporulate and become infectious, so direct human-to-human transmission is generally not the concern. Transmission occurs from fecal contamination of food or water. Fresh fruit and vegetables are the most commonly implicated sources in food borne outbreaks; the food gets contaminated when they’re picked by hand, and they’re then often eaten raw (and unfortunately not always washed well).  In a typical year there would be about 3000 cases of cyclosporiasis reported in the US (and about 300 cases annually in Canada). Fortunately the mortality rate for this disease is very low, but illness can last for weeks even with proper treatment.

With so many people getting sick, it’s important to consider whether there are also risks to – or from – pets and other animals.

Exposure of dogs and cats to Cyclospora

It’s likely that dogs and cats are being exposed to this parasite as well, but exposure is likely uncommon ad fairly low level when it occurs. Currently the outbreak investigation is focused on a source of contaminated iceberg lettuce that was produced in central Mexico. While this iceberg lettuce may not be the only source, it seems to be the main one at this point. Presumably very few dogs (and even fewer cats) are fed lettuce, so that limits their exposure. Indirect exposure to the parasite through contaminated hands or surfaces, or cross-contamination of other foods that the animal might eat is always a potential risk, but the overall risk of exposure for most pets is presumably very low.

What would happen if a dog or cat ingests Cyclospora?

After exposure, a few different scenarios can occur:

  1. The animal gets sick, including shedding the parasite in its feces
  2. The animal doesn’t get sick, but has a subclinical infection and sheds the parasite in its feces
  3. The animal doesn’t get sick or infected, but the parasite passes intact through the animal’s intestinal tract and winds up in the feces (at very low level)
  4. The animal doesn’t get sick or infected, and the parasite doesn’t make it through the intestinal tract to the feces either.

Some foodborne pathogens can cause disease in a wide range of species (e.g. Salmonella). Others, like Cyclospora cayetanensis, are much more limited in terms of species they can infect. Investigations of both experimental infection and potential natural infection have concluded that it’s predominantly or almost exclusively a human parasite. Experimental infection has been attempted in dogs without success; while that’s not a guarantee that dogs can’t ever be infected (experimental infection models are never perfect), it’s certainly supportive of this parasite not being clinically relevant in dogs. Cyclospora cayetanensis DNA or Cyclospora-like oocysts have occasionally been detected in dog feces, but these findings have not been associated with clinical disease and most likely reflect simple transit through the intestinal tract rather than true intestinal infection. Very limited information is available for cats, but there is currently no evidence that cats develop clinical infection or serve as an important reservoir either. So while we can’t say with 100% certainty that dogs and cats can’t ever be infected with this parasite, the odds are very low.

Can dogs and cats spread Cyclospora to people, even if they don’t get sick? 

Living with a dog was a statistically significant risk factor for disease in one Cyclospora outbreak in the late 1990s, but there are lots of potential confounders, so this doesn’t provide any real evidence that dogs themselves are the risk. Theoretically, we can’t rule out a dog ingesting the parasite, passing it in their feces, and then a person in the household getting infected through fecal-oral exposure, but that’s a stretch. Even if viable oocysts were passed in the dog’s feces, they would still require time in the environment to sporulate before becoming infectious, so basic routine practices like cleaning up dog feces (poop and scoop) and hand washing would reduce the risk even further. Cyclospora is very low on the list of things I’d be concerned about in any random fecal sample from a dog. The risk of exposure for people through contaminated produce is much higher.

If there’s active Cyclospora infection in a person in the household, I’d pay close attention to avoiding fecal-oral exposure to dog feces, but I’d say the same thing for any household with a dog.

There is currently no evidence that dogs or cats play a meaningful role in the epidemiology of human cyclosporiasis. If there is Cyclospora in your household, focus on preventing fecal contamination and food borne exposure. Standard hygiene and feces-handling practices are always important, but there is no reason to be concerned about pets becoming ill or serving as significant sources of infection for people.

Here’s a screenshot of a quick but useful New World Screwworm (NWS) resource for animal shelters that was developed by University of Wisconsin-Madison Shelter Medicine Program in collaboration with a number of shelter organizations and shelter medicine experts. With the continued re-emergence and spread of NWS from Central America back into the southern US, it’s an issue of importance for animal owners and veterinarians in those areas, and for those who take care of animals that are moved from those areas. Domestic animal cases in the US (including cattle sheep, goats and three dogs) have been limited so far, but will no doubt continue to increase, at least in the short term. Basic awareness is a critical control tool to stop further spread.

For the original post (and larger print) use this link: Humane Prevention and Management of New World Screwworm (NWS) in Shelter and Rescue Dogs and Cats

test

Last year around this time, were were talking a lot about a cluster of cases of Rocky Mountain Spotted Fever (RMSF) in dogs (and people) associated with Long Point, Ontario. Prior to 2025, we didn’t consider this tickborne disease to be established in the province, but then we detected multiple confirmed cases in dogs, two confirmed cases in humans, and ticks carrying the causative organism (Rickettsia rickettsii) all in (or linked to) Long Point. We assumed we’d see more in 2026 again, but the only way to know was to wait and see.

Because RMSF is not a reportable or notifiable disease in Ontario, we only find out about cases through people, physicians, veterinarians and others who let us know about them, or who ask for advice about suspicious cases. Even if we knew about all the diagnosed cases in the province, we’d still only be getting a partial picture, because for an infected dog to get diagnosed and for us to hear about it, a bunch of things need to happen:

  • the dog has to get sick (not all of them do)
  • the dog has to be taken to a veterinarian
  • the veterinarian needs to consider RMSF (which could be easily missed, since this is an emerging disease in our area)
  • the owner needs to agree to (and pay for) testing
  • the test has to be positive
  • someone needs to pick up the phone or send an email to let us know

There are lots of places along that cascade where we can lose a case, so the ones we hear about are no doubt only a subset of the cases that actually occur.

This year (2026), as part of an Ontario Animal Health Network (OAHN) project, we are testing blood samples from dogs in the Long Point area to see if any/some/many have antibodies against R. rickettsii, to help give us an idea of how much this pathogen may be circulating under the radar. Stay tuned for more updates later in the summer!

We’re also focusing on passive surveillance, i.e. encouraging people to let us know about any confirmed or suspicious cases of RMSF in dogs, and collecting as much information about each case as we can. I’m a bit surprised I haven’t heard about more cases so far, but there has been at least one positive dog that visited Long Point regularly. Long Point is a vacation home area, which complicates surveillance because many dogs may visit the area and potentially get infected, but they get sick back at home and go to their local veterinarian, who may not be aware of the risk for RMSF, especially if they don’t ask about travel history. I’ve also heard that there were other sick dogs in the area, but that’s getting into anecdotes (which can be useful but may also be questionable).

What we can say is: RMSF is back again (or more accurately, it hasn’t gone away, so we’re seeing seasonal cases again, as would be expected). I think we can confidently say there’s ongoing risk to dogs and people that live at or visit Long Point. We can also confidently say that the risk should extend to adjacent areas, but we don’t know how far that risk extends. Had all the dogs around Long Point been seronegative (no antibodies against R. rickettsii), then testing dogs from farther away would have been pretty low yield. But as we learn more about what’s going on at Long Point, we can better design our next surveillance steps.

In the interim, tick avoidance remains paramount, particularly in Long Point, including avoiding high-risk tick environments (e.g. long grass), doing diligent tick checks (daily – for dogs and people), and keeping dogs on a good tick preventive medication (which is helpful but does not guarantee full protection from RMSF).

test

This post is going to be long a long one, as treatment of Brucella canis infection in dogs is a complicated, and there are a lot of important associated issues to mention. TL;DR – I’m not a fan of the hard line “thou shalt euthanize or lock away a dog with B. canis.” I want to prevent disease spread, but we also need to consider the bigger picture and keep things in perspective and the proper context.

To treat or not to treat? That is the question. In some areas, it’s recommended to euthanize dogs that test positive for B. canis, or to strictly isolate them from other dogs and people. This approach is based mostly on what’s been recommended historically out of an abundance of caution (because a long time ago “someone” said we should do it this way) and unsubstantiated fear, rather than hard evidence of risk. Nonetheless, there are some valid issues to discuss.

Personally, I have no problem treating infected dogs if the owner is aware of the potential risks and all the unknowns. Apart from outbreaks and surveillance testing, I get involved with probably about ten new B. canis cases a year. Considering I deal mainly with weird infectious disease situations, that indicates it’s a fairly uncommon occurrence (or at least uncommonly diagnosed), but it’s common enough that we need to have a plan for handling these cases.

It’s hard to know how common B. canis infection really is in the general dog population. Studies of specific populations like dogs in shelters have reported highly variable results, with some reporting over 50% of dogs affected. I suspect that the prevalence in in the general dog population is probably less than 1%, but 1% of millions of dogs is still a lot of infected dogs.

  • Dogs from shelters, imported dogs and dogs from questionable breeding facilities tend to be higher risk. We did a study of B. canis in commercial dog breeding kennels in Ontario in 2019 and 127/1080 (11.8%) of dogs were infected. That was a biased population, as it included some kennels where there were known reproductive issues or potential for exposure to other infected kennels, but it’s still relevant. We found over 100 infected breeding dogs, having tested only a subset of the breeding dogs in that area.

So, we have to assume there were (and probably still are) hundreds of infected dogs out there being bred. That means there’s presumably lots of human exposure, among breeding kennel owners and staff, and among those who adopt dogs from them. Yet, human infections are very rare and most infectious disease physicians I have talked to have not seen one.

Infection in dogs can cause diseases such as diskospondylitis (which causes back and neck pain) but most infected dogs have no signs of disease and probably never will. That’s good for the dog, but it means we can’t tell who’s likely infected just by looking for signs of illness.

The actual risk to people from exposure to a dog with B. canis is low. There’s presumably a fraction of human infections that go undiagnosed or get misdiagnosed, but even so the number of human infections is likely really low. It’s also treatable in people.  Based on a recent scoping review of B. canis cases in humans (Weese & Weese 2025), the literature is quite sparse. We found reports of infections in only 68 people, and 20% of cases with a known origin were lab-acquired, not from dogs. Signs of disease in these cases were variable; while illness can be severe, fever, malaise, fatigue and chills were most common, and no mortalities were reported. Contact with breeding dogs, especially contact with fetuses or fluids after a dog had aborted from B. canis infection, was an important source of exposure. Since then, there was a case report about an infection in a veterinarian in the UK, presented at the 2026 ESCMID Global conference. The likely source in that case was contact with imported breeding dogs. The veterinarian developed forearm blisters, severe back pain, joint pain, and then fever, headache, chest pain, rash and conjunctivitis, but was eventually diagnosed and successfully treated.

I’m definitely not dismissing the risk of disease from B. canis. I don’t want it and don’t want anyone else to get it either. I want to prevent infections in people and dogs. But using an extremely strict approach might not be all that helpful, and could potentially be detrimental if it discourages testing (more on this below).

Risk of disease and transmission varies depending on the type of dog. Consideration of the type of dog involved is important, but often gets overlooked. The highest risk is contact with female dogs that have recently aborted, fetuses and fluids from an abortion, and newborns and maternal fluids/tissues from infected dogs after giving birth to live puppies. This situation doesn’t apply to most households. The risks from most other dogs are likely low, but they aren’t zero; transmission from a sexually immature dog has been reported (once, from a puppy). Contact with urine is probably next on the list of risky fluids, but it’s still probably low risk overall. Theoretically blood and saliva could pose a transmission risk, but those are likely even lower risk.

Another way we can try to assess the dog-to-human transmission risk is to look at the dog-to-dog transmission risk. I’ve been involved in quite a few cases where someone who already owned one or more dogs adopted a new dog that was subsequently found to be infected with B. canis. Almost always, the new dog was already spayed/neutered, or this was done shortly after adoption before the infection was identified. We have often tested the other dogs in the household in these situations and I don’t think we’ve ever found a positive housemate dog, even after the animals have lived together for years in some cases. That doesn’t mean there’s no risk of transmission, but it provides further support for the limited risk in normal household situations.

When in comes to treatment of B. canis in dogs, there are actually two components to consider: treatment of disease versus elimination of the bacterium.

In dogs that actually get sick from the infection, we usually get a good clinical response to antimicrobial therapy. Certain infections, like diskospondylitis, may take a long time to respond, but they usually do eventually.

Because B. canis is adapted to live in dogs, total eradication of the bacterium is tougher; it requires even longer term antimicrobial therapy, and it’s hard to know if/when the B. canis has been eliminated. Typically, I recommend monitoring dogs on treatment serologically, which measures the amount of antibodies against B. canis in the bloodstream. I aim to treat until the dog has no detectable antibodies anymore (seronegative), realizing that that’s likely over-treating, because the antibodies take a while to disappear even after the infection is gone, but unfortunately it’s the only readily available monitoring tool we have. I anticipate at least 3 months of antimicrobial therapy to start, but it usually takes longer than this. Most of the recurrences that I have seen have been in dogs treated with only a month or two of antimicrobials without any serological monitoring. Unfortunately we don’t have great data on overall response to this treatment regimen, since not many dogs are treated and treatment and monitoring are variable. In my anecdotal experience, most dogs will eventually become seronegative and stay seronegative (assuming they don’t get re-exposed), but I still can’t guarantee complete eradication of the bacterium, so the infection could still recrudesce.

So B. canis is widely distributed in dogs and uncommonly causes disease. It probably doesn’t crack the top 10 list of zoonotic disease risks from dogs. So is it really that big a deal? I still wouldn’t dismiss the risk, because even though B. canis infection in people is rare and treatable, it’s still something we want to prevent. The other issue, though, and the one that prompted this post, is the variability in response to an infected dog, which is sometimes euthanasia – and that’s a very big deal.

  • The question shouldn’t be “does a B. canis infected dog pose a risk of B. canis transmission to people?” because the answer to that is obviously “yes, but a very low risk.”
  • The better questions are “Does a dog with B. canis pose a significantly greater overall health risk to people than any random dog?” or perhaps even more importantly “Does a dog with B. canis that has been diagnosed and is being properly managed pose a significantly greater overall health risk to people than any random dog?” I would argue that the overall health risk to people from a well-managed infected dog is almost identical to that of any other random dog that may or may not be infected, but is not being managed for it. The risk may even be lower in a managed dog because we know the dog is getting regular veterinary care, so other potential problems can be flagged.

We often get tunnel vision on one risk, while ignoring other greater risks. Every dog is harbouring multiple bacteria that can cause serious disease in people under the right circumstances. We don’t have any “zero risk” dogs, we have higher and lower risk dogs, people and situations, but anyone can get a zoonotic infection from any dog. More people die every year from Capnocytophaga canimorsus infection from dogs (a bacterium that’s present in the mouth of pretty much every dog) than have probably ever died from B. canis infection Pasteurella (which is also present in the mouth of most dogs and cats) causes many more infections than B. canis ever will. I could write a few pages on pathogen threats that are much more common than B. canis. That doesn’t mean B. canis isn’t important, but we should think about the big picture and the logic (or lack thereof) of an approach as drastic as euthanasia or long term isolation.

More testing of dogs for B. canis would be useful. Ideally, we’d test all dogs that came from a kennel of unknown B. canis status, so we could understand and mitigate the risk of spread. However, harsh draconian consequences disincentivize testing:

  • “Cute puppy. Since we’re not sure about where he came from, we should test him for Brucella canis. But if he’s positive, we’ll have to euthanize him.” How many vets want to make that offer? How many owners would actually agree to testing knowing that?

Disincentivizing testing means we get fewer dogs tested, so we treat fewer infected dogs, and we know less about the infection itself and response to treatment.

At the same time, differential application of severe measures is illogical and ineffective. I have a hard time buying into really strict responses to a known infection, while there’s a much larger number of unknown infections that are simply ignored. It doesn’t make sense to euthanized the dog that responsibly gets tested, while all the other infected dogs and their owners are allowed to simply remain blissfully unaware. If the risk from the known-infected dog can be manged (spay/neuter and treatment), it’s going to be much less than the risk from all those other unknown positives.

So, what’s a veterinarian to do when a dog tests positive for B. canis?

The first step is talking to the dog owner. They need to understand the situation and the options. If they decide to treat, that will involve months of medication and periodic repeated testing. That can get expensive, so they have to be willing to commit. They can’t just give a few days or a few weeks of antibiotics and hope for the best.

If not already done, the dog needs to be spayed or neutered. That will greatly drop the transmission risk to people, and prevents breeding-associated infections, which is the main dog-to-dog route of transmission.

The owner needs to realize the limitations of treatment. We can never guarantee that B. canis has been eradicated from the dog, because the bacterium is notorious for laying dormant in tissues and sometimes infection will recrudesce after treatment. With spaying/neutering, the right antimicrobials, adequate treatment duration and recheck monitoring, the risk of recurrence is probably low, and the transmission risk from a spayed/neutered dog with an early recurrence that is detected through routine testing is probably very low. Yes, there are a lot of probablys here, but that’s the nature of a lot of understudied infectious disease issues.

Treatment is not the right thing to do in every situation. If the owner can’t commit to proper management or has particular concerns about B. canis, rehoming the dog may be a better option. Some rescues will take infected dogs, treat them, and find homes for them with people who understand the situation and risks, which is great. Those are niche rescues and niche adopters, but I think it’s perfectly appropriate, for all the reasons above.

If I suddenly found out that my dog Ozzie was B. canis-positive, I wouldn’t consider euthanizing him. He’s neutered, he’d be treated and he’d be monitored. The risk to my family from B. canis would be exceptionally low, we’d be willing to accept that risk (just like we accept many other risks from owning a dog and cats).  Not everyone does risk math the same, and if someone didn’t want to keep a positive dog, I wouldn’t consider that inappropriate at all. You have to do what works best for your scenario.

What I don’t want to see is unnecessary broad draconian measures and loss of family pets for limited benefit in terms of disease prevention. There’s no right or wrong, but a blanket “you have to get rid of the dog” is the wrong approach to me.

test

UPDATE: June 11, 2026. Nitenpyram has also been given emergency use approval (EUA) for dogs and cats over 2 pounds and 4 weeks of age.

While completely expected given the steady march of New World screwworm (NWS) through Central America and Mexico over the last several years, this horrific parasite finally crossed the US border back into Texas in the last week, and has thus far been found in several beef calves, a goat and one dog from New Mexico. There’s an overview of NWS for companion animal veterinarians recently published in Clinician’s Brief, or check out our previous post NWS: Primer for Canadian Veterinarians for the basics on this nasty bug. This post is about what drugs we have available to help with prevention and treatment of NWS. Remember that the core component of treatment is removal of the maggots as quickly as possible (and ensuring that they do not reach the ground if they are removed or drop out of a wound), but antiparasitic drugs can be helpful as well, especially for the larvae that may be missed or hard to reach, and they can also help prevent infestation before the damage is done.

Compared to other emerging/re-emerging diseases, we actually have quite few options for treating NWS. Data supporting different drugs is variable, and come from a combination of field studies and basic drug information. In the US, the FDA has given conditional approval or Emergency Use Authorization (EUA) to quite a few products for use against NWS in various species. Beyond that, there are also some other products that should work… lack of conditional authorization or EUA in these cases may be more a reflection of some companies not bothering to pursue the process, rather than the product not being efficacious against this parasite.

Among the effective products for dogs and cats, the main drug class is the isoxazolines. This includes lotilaner, afoxalaner, fluralaner and sarolaner. This class has proven efficacy against NWS, and some products now have approval in the US (more on that below).  Lotilaner’s approval is probably based largely on one small study that showed 100% efficacy (i.e. all larvae died) within 24 hours of a single dose in naturally infested dogs after a single dose (Lopes do Vale et al. 2023). Similarly, afoxolaner (Cutolo et al. 2021) and sarolaner (Oliviera et al. 2019) have shown 100% efficacy in their own small field studies.

Here is a list of some of the products that can potentially be used against NWS in dogs and cats, as well as their current approval / authorization status in the US (if applicable):

Conditional FDA Approval for treatment

  • Credelio Quattro CA-1 (lotilaner, moxidectin, praziquantel): dogs

Emergency Use Authorization for treatment

  • Credelio (lotilaner): dogs and cats
  • NexGard (afoxolaner): dogs
  • NexGard COMBO (esafoxolaner, eprinomectin, praziquantel): cats

Other

  • Nitenpyram also had 100% efficacy within 24 hours of a standard dose in a small study of naturally infested dogs (Correia et al. 2010).
  • Spinosad might be useful, but might be less effective, based on one small study that reported 80% efficacy in treating dogs (Oliveira et al. 2018). I’m wary of comparing all these studies since they are all quite small and heterogenous, but at the same time I wouldn’t reach for spinosad first if there’s no advantage to this product over other available options that might be more effective.
  • Fluralaner (the active ingredient in Bravecto) has Conditional Approval in the US for prevention and treatment of NWS in cattle, but it’s reasonable to assume it would be similarly effective in dogs and cats, in which it’s used at a similar dosage.

Macrocyclic lactones are another class of antiparasitic drugs used as preventatives in dogs and cats. Examples include doramectin, which has conditional approval in the US for prevention and treatment of NWS in cattle, and ivermectin which has EUA for prevention in cattle. It is critical to note that the doses of these drugs used for heartworm prevention in dogs and cats are very low (e.g.4-6 ug/kg) compared to what is used for prevention and treatment of NWS in cattle (200 ug/kg). So while ivermectin, moxidectin, selamectin and milbemycin might be useful at high doses, we cannot assume that dogs/cats getting products like Heartgard, Inteceptor, Sentinel, Milbemax, Proheart and Advantage Multi for routine parasite prevention have any protection against NWS.

Imidacloprid, which is found in products including Advantage II, Advantage Multi, Seresto collars and K9 Advantix II, is also unlikely to have any significant effect on NWS based on its mechanism of action and the low drug levels found in tissue (where the maggots are found) versus in the skin (where this drug is most effective).

None of these products have approval for prevention of NWS in dogs and cats. That’s likely more of a labelling/regulatory issue rather than a true issue with efficacy. Treatment data are simply easier to get: You find some infested animals, treat them, and see what happens. Demonstrating a preventative effect is tougher, especially in the field: You need to put a bunch of animals on the preventive, and then compare the infestation rate in those animals compared to untreated control animals. For a relatively rare disease (at least in this part of the world), you’d need a massive number of dogs or cats to show a preventative effect. Alternatively, an experimental model could be used whereby animals are given preventives or placebo and then deliberately exposed to the NWS flies, but a study like that is a lot more work, a lot more money, and there’s far less appetite for unnecessary use of animal infection models (especially for a nasty parasite like this). Given what we know about the parasite, the drugs and the disease, I think it’s fair to assume that a drug that works for treatment will be useful for prevention. Although I’d never guarantee complete protection, I’d be quite confident in this group of drugs.

The bad news for the southern US is that NWS has returned and will probably be around for a while, as eradication efforts will take time.

The good news is that, unlike for many other emerging or re-emerging diseases, we have effective treatments already available.

We don’t have any approved products in Canada for treatment or prevention of NWS, and I suspect there won’t be much push from drug companies to get any such approval or authorization, since there’s no threat of NWS becoming an endemic issue here. However, we can approach the management of travelling or imported infected dogs in the same way as it’s done in other countries. The treatments in Canada would be extra-label, but that is not uncommon for many conditions we treat, especially when they are rare in Canada.

test

As for a lot (or most) conditions we treat with antimicrobials in dogs and cats, data on how best to use antimicrobials before, during and after dental procedures are sparse. We know that antimicrobials are used a lot for canine and feline dental procedures (Weese et al. 2023), and that a large percentage of said use is probably unnecessary. The same is true in human dentistry; even though they have more data and more guidelines, they still struggle with overuse and poor compliance with best practices.

Part of guideline development is assessing the available evidence. Even when we know evidence is lacking, we still want to assess the issue carefully to makes sure there isn’t more evidence than we thought, and to critically assess what little data may be present to help highlight knowledge gaps.

We recently completed a systematic review of infectious complications of dental procedures in dogs and cats, and the impact of antimicrobial prophylaxis (Weese et al. 2026). Specifically, we addressed a series of questions about use of antimicrobials for routine dental procedures, looking at both peri-procedural (i.e. shortly before and up to 24 hours after the procedure) and post-procedural (i.e. more than 24h after the procedure) antimicrobial administration.

  • Unsurprisingly, we found that data were really limited. That’s important itself.  It’s a gap that needs to be filled.
  • We also found that studies rarely looked at (and were rarely well designed to look at) the key outcomes we care about, i.e. clinical disease. We settled on looking at sepsis, infective endocarditis, local tissue infection and adverse events as the critical outcomes.
  • After some debate, we also included bacteremia (bacteria in the bloodstream) as an outcome of interest, even though it’s not a clinical outcome. Bacteremia is a state that may or may not be associated with disease. People and animals are regularly bacteremic and typically have no consequences from it at all. However, bacteremia can progress to infection in rare circumstances. We wouldn’t normally use a non-clinical outcome like that (especially one that is as common as bacteremia), but most of the comparative studies in dogs and cats focused on bacteremia. So we included it, recognizing the limitations.

The systematic review highlighted the commonness of bacteremia associated with dental procedures, as shown in the figure below of the meta-analysis. The proportion column is the prevalence of bacteremia associated with dental procedures in each study, and the pooled prevalence across studies was 47%.  

These studies didn’t report clinical consequences of bacteremia, but when you consider how many dogs and cats undergo dental procedures, the really high bacteremia rates and the really low apparent incidence of dogs and cats getting sick after their procedures, you can understand how bacteremia (seemingly not a big deal) is not a good proxy for clinical disease (which is a big deal).

To investigate this further, we asked whether dental-associated bacteremia is associated with sepsis, endocarditis or other infectious consequences. Only two studies looked at this. Both were small (total of 22 dogs with bacteremia and 11 without) and no infectious consequences or disease were noted in these animals. The data were so limited we couldn’t evaluate them further or draw any conclusions.

We then asked “In dogs and cats undergoing dental procedures, does peri-­procedural administration of antimicrobials reduce the incidence of endocarditis, sepsis, bacteremia or other infectious complications?” But no studies looked at these clinical outcomes. Only two small studies looked at the effects on the incidence of bacteremia, and the meta-analysis (see below) didn’t suggest there was any beneficial effect. The certainty of evidence was really low of course, based on a variety of factors.

That doesn’t mean we can say antimicrobials for dental procedures don’t have any effect, it just means we have nothing suggesting they do. If we look at human dentistry where there are stronger data, antimicrobials are not indicated for these procedures, apart from in certain high risk patients. Our lack of identified effect has to be tempered because the data are weak, but there’s also nothing suggesting antimicrobials are needed (or useful) for routine veterinary dental procedures in most patients.

If there’s no evidence that peri-procedural antimicrobials are useful, there’s even less reason to think that post-procedural antimicrobials are useful, yet this is commonly done in veterinary and human dentistry. We aimed to address the question “In dogs and cats undergoing dental procedures, does post-­procedural administration of antimicrobials reduce the incidence of endocarditis, sepsis, bacteraemia or other infectious complications?” There were no data for this at all. If we look at data and guidelines from human dentistry, routine post-procedural antimicrobials are not typically recommended, or useful.

Did this review actually tell us anything? Yes and no. I think we ended up with the conclusions we were expecting, i.e. bacteremia is commonly associated with dental procedures, there’s no evidence that antimicrobials are useful for routine dental procedures, and the data are really limited.

The fact that we have no supporting evidence for routine use of antimicorbials for dental procedures was an important finding in this case. When we have an intervention (i.e. antimicrobial use) that has potential harms (e.g. cost, adverse events like diarrhea, selection for antimicrobial resistance, administration challenges, pain from administering a pill to an animal with a sore mouth), the default ought to be not to use it unless we have compelling data or a plausible reason to think that the benefits outweigh the harms. For the routine veterinary dental procedures, we have no such data or reason to justify routine antimicrobial use.