Bad dog! Not on the bed!

Veterinary medicine is a bit of an ugly step-child in the medical world. We practice our arts in a murky corner, struggling to keep up with the scientific knowledge in human healthcare and limited by various ethical, financial, and technical constraints. Never mind the limited time and financial resources, never mind the sharp teeth and claws unleashed on us by our patients; one of the hardest things about this job is trying to be a source of reliable information for animal owners. Clinical research in veterinary medicine makes up a tiny fraction of the overall medical literature. We rely on experience and extrapolate evidence from one species of bipedal ape perhaps more than we should. Often we resort to downright superstition. We are research poor.*

Take the case of urinary incontinence (UI) in female dogs. This common condition is a great illustration of the current disparity between how we practice veterinary medicine and what we’ve actually proven scientifically. Most veterinarians will tell you that if your spayed dog wets herself when she’s asleep, she has what’s colloquially known as “spay incontinence”. For years we’ve assumed that spaying female dogs (in some cases) leads to an inability of their urethra to completely seal off, resulting in UI. We’ve been spaying dogs for many decades, and long assumed that UI was one of those tradeoffs for an individual dog that is the price of spaying (among the many benefits of spaying, perhaps the most important is that Bob Barker told you to do it, and you don’t want to be on his bad side).

Most cases of UI are pretty medically benign, and most dogs do well on oral medications, though complications do (rarely) occur. However, these “accidents” can be a source of major frustration for pet owners, and some will actually relinquish their animals to shelters or have them euthanized. So it’s a minor problem for most dogs, but in many individual cases it matters quite a bit. You would think such a common issue that impacts the lives of millions of animals and people over many years would be well worked out.

Uhh, yeah…not so much.

There have been a smattering of research articles over the years but not what you would call a robust buffet of data. In recent years, we as a profession have been a bit concerned about some of the long term health effects of spaying and neutering, and what effect the age of the dog when she’s spayed is as well (summarized nicely here). There has been a flurry of activity to try and figure out what the costs actually are. Better late than never, right?! Urinary incontinence, being such a common problem, finally got a systematic review in 2012, evaluating all of the evidence that had been published on the issue of what effect spaying has on UI in female dogs. It was worryingly slim; only seven eligible articles, four of them had to be thrown out due to high risk of bias, and then mixed results from the last three. They ultimately had to conclude that at best:

The evidence for a causal relationship between neutering and urinary incontinence is weak, although there is some evidence of an association.

Hmmmm. That, like waking up to a urine soaked bed, is a bit awkward. Here we have veterinary dogma, the link between spaying and UI, being exposed as resting on evidence weaker than a leaky urethra. Only one of the eligible studies actually directly compared spayed and unspayed dogs. On the question of how important age at spay is, there were conflicting results, with two studies that said there was no effect and another study that said there was. Each study had only a few hundred dogs in it, and measured different variables in different ways. For example, the study that found the association between age at spay and urinary incontinence looked at dogs spayed as young as six weeks, whereas the other studies’ dogs weren’t spayed until 16 weeks. And it wasn’t certain that the age at spay was even recorded accurately.

What a mess. There have been a few more studies since this review that have given us some information about prevalence and possible risk factors, but each one of these is also limited by a small sample size among other factors. We have a very small number of studies with high risk for bias and a range of results. This is our best evidence for a condition that affects probably hundreds of thousands if not millions of dogs, and is one of the (we think) most common long-term complications of spaying. This is a problem that we should have well figured out by now, and it turns out all we have a hodgepodge of papers with various degrees of bias.

Typically in clinical veterinary research, we as day to day practitioners hope that a handful of academics in tertiary referral veterinary hospitals will do their thing and churn out the science to better inform us. But this top down system often puts conflicting interest at odds. The types of conditions that are important to someone who is implanting fluoroscopic-guided pacemakers in miniature donkeys** are pretty different from the concerns of the family vet trying to get Mrs. Fuzzybutt from peeing on the sofa. We just haven’t had good real world data for problems like these. Urinary incontinence, although very common, isn’t a disease of particular severity, which may be why it is so underrepresented in the literature.

Episode IV: A New Hope

Fortunately, we have a new way of looking at things. Dan O’Neill, a veterinarian and epidemiologist at the Royal Veterinary College in the UK, and others have been hard at work on a project called VetCompass. It works completely differently than our typical research models; rather than small groups of animals in specialty clinics, it collects huge amounts of data from primary care clinics and analyzes it together.

The system works by skimming electronic records from clinics who have volunteered their records, and churning out information on patient demographics and clinical information. This generates a huge amount of data that can be analyzed for patterns. This is standard epidemiologic research, which is extremely valuable and also extremely uncommon in veterinary medicine. A combination of lack of funding, relatively recent adoption of electronic medical records in veterinary medicine, and technological constraints have limited large-scale epidemiology in companion animal research. VetCompass is making up for lost time and investigating a lot of problems; in the last ten years the project has yielded over 40 peer reviewed studies, with more than a dozen projects currently in the works.

So how does looking at data this way help us out?

Back to bed wetting Dobermans for a minute. In 2017, the team published a paper after analyzing the VetCompass data for UI in female dogs in the UK. In contrast to the several thousand dogs in the studies on which our previous clinical knowledge about UI had been based, VetCompass was able to look at a dataset of over 100,000 dogs. After sifting through everything, they were able to compare rates of UI for different variables in female dogs. They had data on breed, spay status, body weight, and breed-relative body weight (an estimate for overweightness).

There were a lot of helpful and interesting findings from the analysis. One of the important findings was that the prevalence rate for UI was about 3%, which is likely much more realistic than previous studies rates of anywhere from 2 to 16%. Several risk factors were also identified. Breed mattered quite a bit, with Irish Setters having an eight-fold odds ratio, and other breeds such as Jack Russell Terriers having decreased odds. They also found that bigger sized dogs and dogs who were overweight had a significantly increased risk.

Did it pay to spay?

The big question of what spaying has to do with UI was, of course, one of the big areas of investigation. The data did show that spayed dogs are over twice as likely to be incontinent as unspayed dogs. This convergence of findings from the previous, smaller studies with this larger one strengthens the connection between UI and spaying, although it is less than the eight-fold increase that had previously been reported.

One of the limitations of the data is that spay status is often counted as a binary “yes/no”. If the age of the dog when she’s spayed does in fact play a significant role in the development of UI, it’s possible that there might be no difference in rates of UI between dogs spayed after a certain age, and unspayed dogs. We might be able to keep millions of dog beds dry by spaying dogs a little bit later in life. Alas, this question has still not been answered.

There are a number of reasons why this seemingly simple problem has been so hard to solve. If we had clear data on when each dog was spayed and when/if they had been through heat ever, it could help. But different breeds and sizes of dogs develop at different speeds, so even if two dogs are spayed at six months of age, one might be expected to be physically and reproductively mature, while the other might not. Additionally, UI often manifests years after a spay, so it’s not clear how the relationship between spay status and UI risk changes over time. In other words, it’s complicated, there are probably a number of contributing factors, and we haven’t looked hard enough yet. As I was preparing this post, however, the VetCompass team published results from another study exploring this exact question. They found a similar overall prevalence rate of UI, and similar increased risk for being spayed overall, but when it came to the spay age question, they wrote “No clear association was identified between age at spay and UI diagnosis.”

This does seem to shift the weight toward the side of spay age not being related to whether or not UI develops, and is more evidence that spaying is a more clear long term risk for UI.

Conclusion: Urine for a golden stream of big data research

In the case of UI, our clinical suspicion that spaying causes or predisposes dogs to leak is probably correct, but it’s a bit mind boggling this link has been simultaneously taken for granted and unproven for such a long time. VetCompass has come on the scene and yielded some very practical information. This is important for dog owners and veterinarians when making decisions and recommendations about when to spay.

Science based veterinary medicine has been plagued by a lack of good epidemiologic studies. We are pretty far behind our single-primate-species-treating colleagues, but projects like VetCompass are a wonderful breath of fresh air. Hopefully, more projects like it can harness the power of these large datasets to look at old problems from new perspectives. We have a lot of problems that stand to benefit from this type of research, and it’s exciting to see the projects that are underway already. This type of approach is not a silver bullet however, as has certain limitations of its own (such as dependence on the quality of medical records). But it is part of a comprehensive approach to medical research that can greatly improve the quality of medicine.

* While there are literally hundreds of peer reviewed human medical journals, there are only a couple dozen in veterinary medicine. I want to state very clearly that I have incredible respect and love for anyone who does clinical veterinary research and we’d be much worse off without them. I just wish there were a hundred times as many of them.

**No joke, the first case I was involved with on my clinical year.


  • Greg Bishop is a veterinarian in Oregon who works mainly with dogs and cats. As a fan of the SBM blog, he sees the enormous amount of bad science and information in the field of veterinary medicine as an opportunity, not a problem! But also a problem.

Posted by Greg Bishop

Greg Bishop is a veterinarian in Oregon who works mainly with dogs and cats. As a fan of the SBM blog, he sees the enormous amount of bad science and information in the field of veterinary medicine as an opportunity, not a problem! But also a problem.